In view of the complementary advantages of different focus images for the same target and the problems of unclear focus, blurred edge and ghosting in the existing multi focus image fusion algorithm, a multi-focus image fusion algorithm based on Parameter Adaptive Pulse Coupled Neural Network (PAPCNN) and Convolutional Sparse Representation(CSR) is introduced. Based on the decomposition of high-frequency and low-frequency coefficients by Non-Subsampled Shearlet Transform(NSST), the low-frequency coefficients are fused by CSR, and the high-frequency coefficients are fused by a Parameter Adaptive PCNN(PAPCNN) algorithm. The implicit function β in PAPCNN is improved to achieve better fusion effect. The experimental results show that the proposed method solves the problems of the traditional PCNN algorithm, such as the difficulty of setting parameters in image fusion and the poor performance of the traditional sparse representation in detail preservation. It has greater advantages in visual effect and objective indicators compared with the existing mainstream fusion methods.