在传统的模糊推理方法中,如果出现模糊规则库稀疏的情况,模糊推理就得不出正确的推理结论.针对这个问题,Kóczy和Hirota提出了一种线性插值推理方法.线性插值推理方法解决了稀疏规则库情况下如何得出推理结论的问题,但是,用这种方法得出的结论有时是不正规的模糊集.本文提出的基于泰勒级数的Kóczy线性插值推理方法,能保证“当模糊规则A1= >B1 ,A2 =>B2 和推理前件A* 是正规的线性隶属函数(三角形或者梯形)时,插值推理结论B* 也是正规的线性隶属函数(三角形或者梯形)”. In the traditional fuzzy reasoning method, if the fuzzy rule base is sparse, fuzzy reasoning can not get the correct reasoning conclusion.For this problem, Kóczy and Hirota propose a linear interpolation inference method.Linear interpolation inference method solves the problem of However, the conclusion drawn by this method is sometimes an irregular fuzzy set.In this paper, the Kóczy linear interpolation reasoning method based on Taylor series can guarantee that "when fuzzy Interpolation Inference Conclusion B * is also a formal linear membership function (triangle or trapezoid) when the rules A1 => B1, A2 => B2 and the antecedent A * are normal linear membership functions (triangles or trapezoids).