At present, Light has become a key factor of affecting face recognition rate, the common method of deal wih illumination were compared in face recognition, experimental results show that the illumination cone method and the approximate light cone light methods have strong robustness when have significant illumination changes. while the other algorithms in the changing illumination cases, the recognition rate has significantly decreased. Deal with three-dimensional face image algorithm, illumination cone algorithm and the spherical harmonic theory have been elaborated. Extensive in-depth efforts have been put into the research toward illumination treatment methods in face recognition. Their strategies, advantages/disadvantages and performances are elaborated. Several promising directions for future research have been suggested.