In order to solve the problem of the poor performance of the traditional microwave resonance method in multi-parameter fitting data processing, a permittivity measurement method based on Back Propagation (BP) Neural Network algorithm is proposed, which introduces the Neural Network algorithm in data processing of microwave resonance method for the first time. In order to verify the effectiveness of this method in measuring permittivity, a microstrip line structure is used as a microwave resonator. It achieves high sensitivity (4.62%) by loading periodically arranged open resonant rings. On this structure, the reflection coefficients S11 of different material samples are simulated as the data of neural network. The amplitude and phase of S11 and resonant frequency f are taken as the input layer of the neural network, respectively. The dielectric constant and dielectric loss are taken as the output to establish the neural network model. The simulated and measured results show that the dielectric constant and dielectric loss calculated by the model are basically consistent with the data provided by the manufacturer. The relative error of the dielectric constant is less than 0.6%, and the error of the dielectric loss is less than 0.0005. Compared with the traditional data processing of microwave resonance method, the introduction of BP neural network algorithm can significantly improve the accuracy of dielectric constant measurement.