The identification of the scientific collaboration network structure using network theory can help to understand the collaboration patterns and behavior mechanisms among scientific individuals. Applying with complex network multi-scale paradigm, the multi-scale network methods, i.e., marco-, mesco-, and micro-scale concepts, algorithms and applications for analyzing the global, module, and motif structure of the scientific collaboration networks were explored. Furthermore, these methods were effectively applied to reveal the multi-scale nework collaboratin patterns and behavior mechanisms of the real scientists’ collaboration network in the field of complex network research. The research may provide hierarchical ideas and methods for detecting, analyzing and managing the collaboration patterns and behavior mechanisms in complex scientific collaboration systems.