The frequency aliasing of the original contourlet transform was investigated, and a new multiscale decomposition scheme was designed to suppress the aliasing effect. Then, by the combination with directional filter banks, an aliasing-free contourlet transform (AFCT) was realized. Finally, a novel image fusion algorithm based on AFCT was proposed. By cal- culating the neighborhood homogeneity measurement (NHM), the transform coefficients can be fused local adaptively. Ex- perimental results demonstrate that the proposed approach outperforms the traditional methods which are based on wavelet transform and contourlet transform in terms of visual quality and quantitative criterion.