In order to enhance the quality of medical image fusion, a novel CT and MRI image fusion algorithm is proposed based on Shift-invariant Shearlet Transform (SIST) and compressed sensing. Firstly, the source CT and MRI images are decomposed by SIST to obtain the low frequency sub-bands and high frequency sub-bands. Then, for the low frequency sub-band coefficients, a fusion rule method combining with a new improved spatial frequency, which improves regional weighted energy and local similarity matched degree, is presented. For high frequency sub-band coefficients, a scheme based on the theory of adaptive 2PCNN-CS is presented. Finally, the fused image is obtained by performing the inverse SIST. The experimental results show that the proposed approach can outperform the conventional CT and MRI images fusion methods in terms of both objective evaluation criteria and visual quality.