鉴于单个预测模型在舰船维修费预测中存在信息源不够广泛的不足,提出用组合预测法进行维修费预测。构建了基于径向基函数(RBF-Radial Basis Function)神经网络的舰船维修费组合预测模型,并用实例进行了验证。结果证明了此方法的有效性。 In view of the insufficiency of information source of single forecasting model in warship maintenance cost forecasting, a forecasting method based on combined forecasting method is proposed. A model of warship maintenance fee combination forecasting based on radial basis function (RBF-Radial Basis Function) neural network is constructed and validated with examples. The result proves the validity of this method.