An efficient simulation procedure for stochastic analysis of nonlinear dynamical systems with multi-degree freedom based on Monte Carlo simulation (MCS) is presented. A simple criterion was established for indicating the importance of each dynamical response sample. According to this criterion, Russian Roulette and Split (RRS) method was applied to deal with the selected samples. The efficiency of this algorithm was much higher than that of direct MCS while the number of response samples in the low probability regions was increased. The result of evaluated examples demonstrates that the methodology presented in this paper has the advantage of being simple in operation, thus applicable to realistic engineering calculation.