Abstract: An adaptive genetic search algorithm is developed for fusion of two spatially registered images of the same scene. In this method, the size of block is defined as chromosome, after crossover and mutation, the global optimal image will be got. Three evaluation criteria such as root mean square error, entropy and mutual information are used on the analysis and effect evaluation of different fused images. Two cases are discussed and extensive experiments demonstrate that in one case the method achieves reconstruction or optimized fusion result to the reference image when the focus objectives are not overlapping blurred, and in another case this method performs better outperforming Laplacian and wavelet methods when the focus objectives are overlapping blurred.