針對四輪式機器人做非完整運動時系統的非完整性的問題，將四輪式機器人運動規劃轉化為非線性控制系統的優化問題。提出了對優化變量進行浮點數編碼的改進遺傳算法，使系統控制精度得到改善。同時將改進的遺傳算法采用最優個體保留策略，設計交叉參數和自適應變異參數，確保算法具有良好的收斂性。通過數字仿真實驗，證明了該方法的對四輪式機器人非完整運動規劃問題具有可操作性。The optimal motion planning problem of four wheeled mobile robot systems, due to the nonholonomic constraints, is usually transformed into an optimal control problem of nonlinear system. The ameliorated genetic algorithm of optimal control for motion planning is presented with encoding the optimized variables with float-point number. The strategy the best individual to be reserved in the ameliorated genetic algorithm operation is applied. With appropriate crossover parameter and self-adaptive aberrance factor, the convergence of the algorithm is improved and the system’s motion precision is leveraged. The numerical simulation indicates the effectiveness of the algorithm for motion control of four wheeled mobile robot systems.