A novel content-based retrieval algorithm is presented for video emotion types recognition in this paper. Considering the fuzzyness and uncertainty of the human's cognitive behavior, we established the relation between low-level video features and high-level emotion types by the unascertained theory. The principle of video features selection and implementing method are introduced in detail in the part of video low-level emotional feature extraction. And the specific process of building the measure matrix and the degree of confidence method are discussed in the part of emotion type determination. Experimental results show that the proposed algorithm is feasible and and effective.