If I have an ellipsoid. If light emmitting in spherical or circular shape from its one focus (focal point), is it possible to reflect maximum of the rays on its other focus (focal point)? what will be the method and how to check/verify it? Is there any other geometry to focus light from one point to another.

Yes, for an ellipsoid of revolution, every ray emerging from one focal point will be reflected to the other focal point. In practice since your source is not an ideal point and there will be some supports or power connections to the source, there will be some rays which reflect back to the source or land on the supports. Are you interested in a mathematical proof or how to check it physically using an ellipsoidal mirror? I feel quite confident that the answer is no, but I'd want to check through the details and write out the proof before saying for sure. If you are interested, let me know and I will gladly explain my line of reasoning --- it is based on Expressing the question as a Pfaff problem, arguing that the solution must be spherically symmetric, reducing to one dimension, where the problem reduces to finding integral curves of a vector field, then noting that the ellipses are all the integral curves.

Thanks a lot for reply. I need not only mathematical proof and also physical(through some computer simulation). If i have a source on the focal point emitting light in spherical shape and the mirror will be ellisoid.

For a proof, see here (corollary to theorem 1): http://planetmath.org/encyclopedia/PropertiesOfEllipse.html

Also consider that the paraboloid is in some sense a degenerate ellipsoid; one focus is "off at infinity".

This proof is only for a 2D ellipse. I need a proof of 3D ellipsoid. Also you discuss reasoning for not other geometry. How i can simulate/program through a computer that when a light ray emit from a focus of ellipsoid will reflect to the other focus of it.