Point-in-polygon (PiP) is a very well-studied computational geometry problem, so there are lots of algorithms and implementations out there that you can use. Searching SO will probably find several you can copy-paste, even.
There's a catch, though—you're dealing with polygons on the surface of the Earth... which is a sphere, not the infinite Euclidean plane that most PiP algorithms expect to work with. (You can, for example, have triangles whose internal angles add up to greater than π radians.) So naively deploying a PiP algorithm will give you incorrect answers for edge cases.
It's probably easiest to use a library that can account for differences between Euclidean and spherical (or, more precisely, Earth-shaped) geometry—that is, a mapping library like MapKit. There are tricks like the one in this SO answer that let you convert a MKPolygon
to a CGPath
through map projection, after which you can use the CGPathContainsPoint
function to test against the flat 2D polygon corresponding to your Earth-surface polygon.
Of course, to do that you'll also need to get your KML file imported to MapKit. Apple has a sample code project illustrating how to do this.