Although using the mobile user interface pattern (MUIP) can increase the efficiency and quality of mobile interface development, the existing pattern search methods have some limitations to provide appropriate patterns for solving a given design problem. In order to find the appropriate patterns, a category theory-based MUIP recommendation (CTR) method is proposed in this paper. First, the MUIP search is represented with a categorical structure so that the design problem can combine with the hierarchical structural MUIP repository closely. Then, to extract more accurate requirement information from the design problem, we present a question-based information extraction method that combines the general architecture for text engineering framework and category theory to extract the design goal, context, and level from the design problem. Finally, based on the categorical structure, we adopt a corresponding search strategy to find MUIPs according to the type of design problem. Some experiments are conducted to evaluate the performance of CTR. The experimental results show that the CTR can provide a high performance in terms of accuracy and ranking when compared with the existing methods. INDEX TERMS Mobile interface development, mobile user interface pattern, recommendation method, category theory, information extraction, categorical representation.