To detect the occlusion region in video accurately, an occlusion region detection approach is proposed for video by fusing multi-feature based on graph cut. Three new occlusion related features named brightness patch match, maximal flow difference and flow residual are proposed based on the information of optical flow and brightness, meanwhile their calculation methods are defined. The feature vector of each pixel is composed of the proposed features and is inputted into the random forest classifier to obtain the occlusion related information about pixels and adjacent pixel pairs. An occlusion detection energy function, which transforms the occlusion detection problem as an optimization one, is constructed by synthesizing the above occlusion related information. An undirected graph is constructed according to the energy function, then the energy function is solved by graph cut theory to gain the final occlusion region detection result. The experimental results show that, compared with the existing advanced methods, the proposed approach has higher accuracy and better real-time performance.