Summary: A Monte Carlo reliability simulation method for a large complex system is proposed based on the reliability block diagram. The relationships of the failure times of various reliability model systems and their component units are given, and the reliability simulation problem of complex systems including bypass models is solved. Choosing reliability simulation data as the sample, the parameter confidence interval as coding space and the reciprocal errors between the empirical distribution function and the acquired distribution function as the fitness function, a point estimation model for age distributions is established based on a genetic algorithm, and the reliability function of complex systems is obtained. Finally, two cases are given to show that this method is more accurate than traditional methods.