This study proposes a medical image fusion algorithm based on improved guided filtering and dual-channel pulse coupled neural networks (PCNN) to solve the problems of blurring edge and complementary information insufficiency in current multimodal medical image fusion. First, medical source images are transformed with a non-subsampled contourlet, and the dual-channel PCNN is used to fuse the low-frequency sub-bands. The sum of the modified Laplacian energy is used as the input of the dual-channel PCNN, and the improved spatial frequency is considered as the connection strength. Then, improved guided filtering is used to fuse the high-frequency sub-bands of the source images. Finally, the fusion of the low-frequency sub-bands and that of the high-frequency sub-bands are inverted by the non-subsampled contourlet transforming to obtain the fused image. Experimental results show that the proposed algorithm effectively retains the characteristic information of the source images and objectively evaluates the mutual information, information entropy, and spatial frequency.