This paper attempts to improve the utilization efficiency of online building data in the age of big data. Focusing on data-driven building design innovation, the authors created an innovation strategy generation method for building design based on the optimization algorithm for dynamic sorting of extension set. Firstly, online building cases were captured by data acquisition software, and processed into a building case library (BCL). Next, the sorting algorithm and operation mechanism of building cases were optimized by setting up a transform mechanism for building problems and query conditions, preliminary screening of building cases based on correlation function, dynamic sorting of building cases based on extension set, etc. Finally, the innovation strategies for building design were automatically generated from the BCL, and the main factors affecting the case sorting were corrected in detailed. The case study shows that our method has high operability, improves the ability of architects to make scientific decisions based on online data, and thus promotes the innovation of computer-aided building design.