Personally, I'd go with Inguarus used by Wikipedia, since the name was basically Scandinavian and the form Igor represents a Slavicized version of that. (I will admit, this is purely a matter of taste.)
Like Draconis, I can't find the name in any Latin sources, but (for what it's worth), it does appear in Leo the Deacon's Byzantine historical work as Ἴγγορ[ος] (Ingor), which looks like a Greek perception of the Scandinavian name. (The context of the form ῎Ιγγορος seems to require a genitive, so I assume the nominative would be Ἴγγορ.)
But, ultimately, it's your choice.
Mostly, it depends what you're using it for.
If someone named Igor travelled back to Ancient Rome, his name would probably be kept the same: Igor is valid Latin, and even looks like a standard agent noun (so it would probably be declined Igor, Igoris, Igorī…).
However, if, say, Rurik had decided to go to Rome instead of Kiev, then the name would have been transmitted from Old Norse as Ingvar. This might have gotten Latinized as Ingvarus, as Wikipedia did, or it might have been an indeclinable foreign noun, or it might have declined like lar (so Ingvaris, Ingvarī…).
For someone like Stravinsky, I would just borrow it as Igor. Unlike other names that have attested Latin predecessors (such as Maria for "Mary"), I don't know of any Latin sources which give a reliable form of this name. So it seems more reasonable to do what English does and just borrow it directly.
(P.S. I'd hoped that some of the old Slavic histories might have talked about Igor son of Rurik in Latin. Unfortunately, all the oldest sources I've found that mention Igor are written in Slavic languages instead of Latin, so they're not much help.)
(P.P.S. I write it Ingvar rather than Inguar because the v was a consonant in this context: compare angvis "snake", which is two syllables an-gvis rather than three *an-gu-is. Basically, after n, gv was a labialized consonant like qv rather than a consonant and a vowel.)