Aiming at the shortcomings of traditional SDG model diagnostic, such as low resolution, slow speed and low efficiency, this paper proposes a fault reasoning method based on fuzzy hierarchical SDG model. Firstly, build system's SDG model, and then layer. Using fuzzy variable to represent node variables, we express qualitative causal relationship by conditional probability table between nodes. Finally we use Bayesian inference and backtracking search to find out fault source candidates, and sort candidate solutions. By using the proposed method, the fault diagnosis of certain type aircraft engine fuel control system is carried out, and the results show that the method can improve the resolution and efficiency of diagnosis, the diagnosis results are accurate and complete.