To comprehensively consider the influence of activity time randomness, machinery failure, collapse, water inflow and other risks on the construction scheme optimization of a diversion tunnel, and realize a reasonable selection of the construction scheme, the construction scheme optimization method of diversion tunnel is proposed considering the influence of risk factors. To begin with, the optimization model of construction scheme considering risk factors is established. In addition, an Levy flight adaptive chaos particle swarm optimization(LFACPSO)algorithm is proposed to solve the model for obtaining the Pareto scheme set. In this algorithm, the Levy flight is adopted to improve particles’ ability of jumping out of the local optimum, then the chaos algorithm is used to initialize particles, and the inertia weight coefficient is adjusted adaptively. Finally, the IFEWPA method is used to select the final scheme considering the experts’ hesitation degree from alternative scheme set. The application shows that, compared with construction scheme simulation schedule neglecting risk factors and planned schedule, the simulation schedule considering risk factors is closer to the actual schedule, which proves the effectiveness of the construction scheme optimization method considering risk factors. Furthermore, LFACPSO algorithm is compared with ACPSO, PSO and NSGA-Ⅱ algorithms, and the results prove the superiority and robustness of LFACPSO algorithm, with the maximum mean value and the minimum standard deviation.