Recently, spectral clustering algorithm has a wide application in pattern recognition and image segmentation. Compared with traditional clustering methods, it can cluster samples in any form feature space and has a global optimal solution. By starting from the equivalence between the spectral clustering and weighted kernel K-means, a spectral clustering algorithm with spatial coherence property of images was proposed. By adding a term of spatial constraints to the objective function of weighted kernel K-means, the algorithm made the minimization of objective function be equivalent to the spectral clustering approximatly. Experimental results show that our proposed algorithm outperforms the traditional spectral clustering in image segmentation.