取代应用原始数据样本构造模糊模型的传统方法 ,提出应用数据变换技术和启发式方法简化模糊建模过程。对于变换后的数据 ,首先通过启发式方法确定模糊 If- Then规则结论部分非模糊单值(即实数 )的初始值 ,然后通过梯度下降学习方法进行精调。该方法不仅模型精度较高 ,而且收敛速度快。仿真实例验证了所提出的模型优于传统的方法 Instead of the traditional method of constructing fuzzy model using raw data samples, the application of data transformation and heuristics is proposed to simplify the process of fuzzy modeling. For the transformed data, firstly, the heuristic method is used to determine the initial value of the unambiguous single value (that is, real number) of the conclusion part of the fuzzy If- Then rule, and then fine-tuned by the gradient descent learning method. This method not only has higher model accuracy but also has faster convergence rate. The simulation results show that the proposed model is superior to the traditional one