中文摘要：

在对城市停车需求分析基础上，提出利用总停车需求、公共停车需求、私人停车需求作为城市停车需求的判别指标。基于BP神经网络对多输入与多输出的数据具有较良好的拟合能力的特点，建立基于BP神经网络的城市停车需求预测模型。以佛山市某停车场为例，分析BP神经网络在停车需求中的适应性，仿真结果表明：BP神经网络预测模型对每组数据的预测相对误差最大为18.80％，最小相对误差为6.21％，符合预测精度要求，具有一定的实际操作性。

英文摘要：

In order to improve the prediction level of urban parking demand ,the total parking demand ,public parking demand and private parking demand were presented as the measurement indexes of urban parking demand .Then ,BP neural network parking demand forecasting model was established according to the BP neural network with the feasible fitting ability for multiple input and multiple output .Foshan as an example to analyze the adaptability of BP neural network in the parking demand forecasting ,the simulation results show that the maximum relative error is 18.80% and the minimum relative error is 6.21% ,the forecasting results meet the prediction accuracy requirements .