The complex electromagnetic environment generated by the deployment of multiple signal sources and radar countermeasures in modern cognitive electronic surveillance methods severely limits the degree of prior information available for effective target identification.In this paper，a dual-view collaborative clustering method based on radar signals is proposed to classify radiation sources，especially in dual-view scenarios.The proposed method iteratively performs unsupervised clustering，cluster label transfer，and dimension reduction through linear discriminant analyses，by which the differences between the clustering results obtained from dual-view scenarios can be distinguished，enabling radiation signal ranking in non-cooperative environments.The experimental results demonstrate that the proposed method can effectively leverage the differences between the basic signal features and intra-pulse characteristics，and enhance the accuracy of cluster-based radiation source sorting.Therefore，the sorting ability of the proposed method has very high practical value.