Because of different imaging mechanism of SAR and visible images，it is difficult to achieve good fusion effect. The SAR image is classified directly. Under the NSCT transform fusion framework，low-frequency weighting coefficient is searched by using genetic algorithm，and the spatial frequency takes big rules to achieve high-frequency coefficients for SAR and optical image fusion. The experimental results show that the proposed method can obtain a better fusion image classification effect，more conducive to subsequent processing such as image recognition.