An important issue in the lifetime data analysis of key parts of equipment is to specify the most appropriate life distribution model. In this paper, weibull distribution as one of the most widely-used reliability models is used for reliability assessment, which is based on the failure data obtained from measurements or simulations. The unary two-parameter or three-parameter weibull model is applied in a system in which the components are independent of each other, and the genetic algorithm(GA) which can search the globally optimal solution is put forward to estimate the unknown parameters. Meanwhile, the Monte Carlo(MC) simulation is adopted to analyze the failure rate under the breakdown and periodic maintenance strategy respectively. Furthermore, so as to model the possibility of multiple failures more vigorously, the mixed weibull model is introduced, and the expectation maximization(EM) algorithm which repeats the iteration process to convergence is presented for the mixed parameters estimation. Finally, the relationship between model and parameters is utilized to assess various reliability characteristic quantities. The results of case studies show that our method can provide the decision-makers with quantitative references to improve the reliability, availability and safety of equipment.