Alhambra 22
Big Ben 23
Brandenburg Gate 20
Chichen Itza 22
CN Tower 14
Colossus 18
Cristo Redentor 21 (Useful to get more policies, although been nerfed so much.)
Eiffel Tower 23
Forbidden Palace 20
Great Firewall 9
Great Library 22
Great Lighthouse 20
Great Mosque of Djenne 20
Great Wall 13
Hagia Sophia 20
Himeji Castle 20
Hanging Gardens 22
Hubble 23
Kremlin 16
Leaning Tower of Pisa 21
Louvre 21
Machu Picchu 20
Neuschwanstein Castle 21
Notre Dame 21
Oracle 23
Pentagon 12 (Upgrade costs are nothing, comparing to the huge GPT in late game)
Petra 23
Porcelain Tower 21
Pyramids 22
Sistine Chapel 20
Statue of Liberty 21
Stonehenge 23
Sydney Opera House 20
Taj Mahal 21
Petra: There are a lot of Wonders I'd like to give the +1 for gameplay reasons - including Hanging Gardens, Macchu Picchu and Mausoleum of Helicarnassus - however of those Petra has to be my number 1 for no other reason that I've visited (and loved) the site recently, and I was thrilled when (a) it was finally added to Civ for the first time, and (b) it proved to be an exceptionally good Wonder with an interesting suite of abilities. Coming with a free amphitheater is a particularly nice thematic nod to the historic site.
Terracotta Army: While I hate the fact that Cristo Redentor is considered a "World Wonder" with a passion, and the Great Firewall even moreso, it's hard to overlook the essential uselessness of the Terracotta Army as an attempt to replace Stonehenge for culture games. It provides less culture than vanilla Stonehenge, it becomes available at an oddly late point in the tech tree to warrant its effect, and even going for culture victory I'd question whether the Great Artist point is better than the very early accumulation of GE points from its predecessor. But above all that, it's just not remotely interesting as an effect - it's just a unique museum without specialist slots you can get two eras earlier. Sure, Angkor Wat is also rather useless, and the site is far too impressive to warrant such a useless effect (which I imagine is also true of the TA, which I've never visited), but I like the Angkor site too much to list that as my least favourite.
EDIT: Oops, seems I'm not following the rules, and Terracotta has already been eliminated anyway. Changed the above figures, also replaced my vote for Terracotta with Great Firewall, which combines being largely useless with being a really stupid idea for a Wonder. And yes, the AI built it in my last game, and no I didn't care in the slightest - as others have reported, by that time the game was won.
Great Library: Probably one of the best early game wonders aside from Petra (but this one will be hyped anyway ).
I like the GL mainly for nostalgia reasons (it's been my beeline Wonder since it was truly powerful in the first two Civ games), but I ran down the Ancient/Classical Wonder list a while back and concluded that, overall, the GL is one of the least useful even if you get it. You have to try for it so quickly that you get a fairly mediocre beaker boost, in effect; over the course of a game the beaker output of the population gain from Hanging Gardens will far exceed the value of the GL. The Library has its uses for very specific tech beelines, but even most of those aren't within a tech of Writing in the G&K tech tree as they were in vanilla. In answer to one question, if the GL wasn't beelined by the AI, I probably still wouldn't get it most of the time - the short-term tech boost you get has to be balanced against getting a library 10 turns or so later than you otherwise would, aside from other considerations.
considering that the Oracle and great Wall both give 3 
