In the traditional workflow-based collaborative design, the difficulties of communication and task coordination among different professional designers lead to low efficiency of product design. In order to solve this problem, the "A Meta-model with Three Levels" data model of complex product and data-driven collaborative design technology of complex product based on data driven were proposed. Firstly, multi-dimensional and multi-granularity data modeling and ontology description were used to complete the information modeling of complex products. Then the semantic retrieval technology based on ontology was used to complete the data subscription of the collaborative design process task. Finally, a complex product task collaboration technology based on data subscription/publishing was implemented. The experimental results show that the data-driven collaborative design technology solves the difficulties of communication and task coordination among different professional designers in the traditional collaborative design process, and achieves a spiral of rise product collaborative design process, thereby improving the efficiency of complex product design.