An active suspension system of vehicles by using the genetic algorithms and neural network controls strategy has been proposed. A half-car four-degree of freedom suspension vibration model was described. Compared with the conventional passive suspension system, the analysis has been done to the system control performance. The analysis of the system response was obtained through the change of the neural network training coefficients, genetic algorithms input functions and the change of velocity. The simulation results indicated that the vehicle vibration could be reduced and the ride comfort is ameliorated by the proposed suspension systems.