Many people ask something for openings here while there are many materials on this website...
https://www.chess.com/article/view/whos-afraid-of-smith-morra-gambit
Another interesting article about gambits in general, though it deals with BDG :
https://www.chess.com/article/view/defending-against-gambits
In any case, it's better to take 2...cxd4 and then, on 3. c3 U could play 3...e6, e5 or 3...Nf6 (as suggested by DeirdreSkye). 
In any case, it's better to take 2...cxd4 and then, on 3. c3 U could play 3...e6, e5 or 3...Nf6 (as suggested by DeirdreSkye).
U forgot 3...d5, reaching another main line in Alapin, 3...g6 (a reliable choice, especially for accelerated dragon players), 3...d6 (often followed by g6), 3...d3 (leading to Maroczi structures if white plays with c4), 3...Nc6 (a transpositional move), and even 3...Qa5 (which avoids cxd4).
The moves u suggest 3...e6 and 3...e5 are worse than those above in my opinion...
In any case, it's better to take 2...cxd4 and then, on 3. c3 U could play 3...e6, e5 or 3...Nf6 (as suggested by DeirdreSkye).
U forgot 3...d5, reaching another main line in Alapin, 3...g6 (a reliable choice, especially for accelerated dragon players), 3...d6 (often followed by g6), 3...d3 (leading to Maroczi structures if white plays with c4), 3...Nc6 (a transpositional move), and even 3...Qa5 (which avoids cxd4).
Ah, yeah, right, these traditional moves are also relatively good, thanks for this correcion. 
Very simple: take everything you can take and then convert the pawn more.
White committed 2 obvious mistakes, 2. d4 and especially 3. c3, which is already a blunder, so black enjoys big advantage.
I can never agree with what some others would claim to transpose to other lines.
No, the best line of play, logically and in terms of concrete variations is to simply accept the gambit.
Here an example game by SF, it understands that line sufficiently well:
The best line of play is to play d6, e6 and a6, getting to some kind of a hedgehog position, so the white pieces can not make use of their better development to penetrate, and then develop, retaining a full central d pawn more, which is a lot.
My lines are good, very good indeed, in the opening and everywhere, it is a pity people treat me in that way.
This is what I do:
And I say, thanks for the offer, but no thanks.
Now lets play some chess.
white gets slowed down a little, because his knight cannot jump to c3 right away. (Its natural square.)
Now fair warning, theory says white should be better, by playing 4.c4, setting up a Marcozy Bind
But the way I figure it, any 1400 or 1600 player that plays 2.d4, isn't going to be comfortable with the white side of a Bind.
No .... there's no hurry for that.
Bxd3 would regain material AND develop a piece at the same time. But I guess there really is no hurry because Black's pawn on d3 is absolutely indefensible and beyond helpless, and White can wait for a while and develop other pieces (say, 3. Nf3) before capturing the pawn.
Very simple: take everything you can take and then convert the pawn more.
Good advice (at least, in this case). Is this the famous Tsvetkov's sustem in his epic fight with Stockfish ?!? 
Esserman played the Smith Morra aginst Anand and Anand preffered to transpose to Alapin.
That was a classical time control game.
You believe you know more than Anand?
That is why he only drew.
Tradewise Gibraltar, Catalan Bay, not very reputable, is it?
Me + SF certainly we know more than Anand.
Especially when we analyse at home.
I bet Anand has not analysed that in-depth.
Esserman played the Smith Morra aginst Anand and Anand preffered to transpose to Alapin.
That was a classical time control game.
You believe you know more than Anand?
That is why he only drew.
Tradewise Gibraltar, Catalan Bay, not very reputable, is it?
Me + SF certainly we know more than Anand.
Especially when we analyse at home.
I bet Anand has not analysed that in-depth.
Sadly for you Anand is an iconic figure of chess and is widely regarded as one of the best players in history, whereas you are, well, Joe Nobody.
The main options you have are to accept the gambit or to transpose it into another line. There is nothing wrong with accepting the gambit (Larry Evans made Smith look silly when he debuted it in the 1972 tournament where Smith lost every game he played the gambit). However, as a Sicilian player you are more likely to see the Alapin, so transposing into it can decrease your preparation. As my club has a couple of Morra-addicts as well as a few Alapin players, I have adopted the approach of transposing. I quickly get into lines where I am familiar with good results. The key is to find a position you are at least somewhat comfortable and work to transpose into that position.
White has some advantage in that line, and I doubt black can substantially improve somewhere.
So that, instead of accepting the gambit and getting nice 70-80cps advantage, black will choose to transpose to some Alapin, whatever that is, and trail behind by 30-40cps.
Why do that?
On the other hand, I agree with you in human OTB games it is very important to know your setups, even if you don't proceed optimally.
In the alternative line Mr. Talparov chooses, after e6(instead of de5) 0-0 Be7 Qe2, SF sees quite some white edge, although the game is very complicated and unhuman-like.
Again, my central backward-maker on e5 makes me recognise white has nice options ahead.
I am sticking to my pattern guns, I just learned about the Smith-Morra and Mexican Attack/Defence, you might want to stick to long opening theory lines that change evaluation every 2 months or so.
19...Qf5 seems like a weird move. Better is probably 19...Qg6!, as played in a high level correspondence games, as well as by some 2600+ GMs. Besides, 7...dxe5 is just one line, "everyone knows" that 7...Nb6 leads to dull equality.
Please stop showing one-minute games by Stockfish as "evidence". I can't understand why Black flung his Rook over to a3 in that last game.
OK, in this critical position Mr. Talparov says white should play Bd2 and black had good play, as white has pawn weaknesses.
Is this so?
SF suggests Qe4 instead, threatening Bd3, and black should weaken the shelter with g7-g6 or something similar.
After that, white should have big advantage.
Again, what makes me think so?
The e5 white central backward-maker that renders the black king shelter weak.
Did we find a novelty on move 13, thanks to SF, that refutes black's play?
Maybe we should communicate that to Anand, and then he might take me for his second.
Is anyone able to refute 13. Qe4?
Again, you might laugh at me, but my lines are very very sound.
Anyone who would consider my book would get much much stronger after some effort to learn the patterns, but people prefer to ignore me and the knowledge the book heralds.
OK, no matter where I post, people start attacking me.
I will not be replying to everyone, of course.
Just refute a single line of mine, with concrete variations, a single one!
Post a position and then say: 'Here you went wrong'.
No one does so, for some reason.
SF at 1 minute playing bad, come on.
The rook, why was it flung to a3?
Well, there is nothing better, what to do, stay on the d file instead?
The rook on a3 is very active.
The point is black's weakened shelter, details will not solve it.
Nb6 dull draw.
Who said that?
My pattern estimation is white always retains some edge.
Again, guys, concrete lines, concrete lines.
Opening theory changes, if Carlsen deviates on move 2 or 3, certainly there is a lot more still to discover.
Why doubt that?
This is an actual Smith-Morra Gambit situation unlike the previous thread.
I'd usually capture the pawn, but smart people would not capture back with the queen and instead play c3, which is the point of the gambit. Does anyone have a good way to counter the gambit?