Resistance measurement is extremely critical in semiconduetor technology. When measuring theresistance of thin films by the traditional four-probe method, nonlinear fiting of the Van der Pauw function isrequired, which is not only time-consuming but also has poor accuracy,In view of this phenomenon, a Backpropagation neural network ( BPNN) neural network model based on the Levenberg-Marquardt (LM)algorithmis proposed. The IM algorithm combines the advantages of the gradient deseent method and the Newtonmethod, quickly approaches the global minimum during the iteration process, and is better than the puregradient descent method in the case of local minimum fall, The weight is adjusted by combining the backpropagation error of the BP neural network to achieve the fitting of complex nonlinear funetions, "The localparameters of the transeendental funetion ( Van der Pauw funetion) containing the inverse hyperbolie cosine arenonlinearly fitted, and a neural network fiuing model with a maximum deviation of 2.08x 109 and a relativestandard deviation of 2. 16 x 10-° is obtained. Compared with the normalized polynomial fitting method, theaccuraey is improved by 99.596.This improved method can greatly improve the stability and accuraey of themeasurement results."The model is applied to the experimental measurement process , which elfectively improvesthe resistivity test accuracy.