A remote sensing image fusion algorithm based on non-subsampling Contourlet transform coupled feature voting mechanism is designed in order to solve the problem of image coefficient fusion. Hue Saturation Value(HSV) transform is adopted to extract brightness components from multi-spectral images. Then the brightness component and panchromatic image are transformed by using the non-downsampling Contourlet transform for calculating the high and low frequency coefficients. The feature voting mechanism is established to fuse low frequency coefficients. The fusion rules of high frequency coefficients are constructed to obtain the fused high frequency coefficients. The fusion coefficients are processed to generate the fusion image. The experimental results show that the fused image of the proposed algorithm has better spectral characteristics and clarity.