To study the fault diagnosis method of many-movement complex electromechanical products, we first decomposed complex electromechanical products according to functions-movement-action(FMA) structural decomposition and established a reliability and fault diagnosis model based on meta-action decomposition and signed directed graph. Then, to improve the ability of fault identification, we brought the meta-action quantitative information of complex electromechanical products into the reliability and fault diagnosis model and diagnosed the fault of meta-action quantitatively by fuzzy theory. The proposed method can improve the fault identification ability of products by monitoring the operating status variables, and provide a fundamental model for fault root cause analysis. Finally, certain machining center was taken as an example to verify the effectiveness of proposed model and analysis method.