The work aims to propose a denoising method of weighing signal based on BP neural network particle filter, in order to effectively filter out the noise signal in the automatic weighing control system and improve the stability and precision of the weighing system. In particle filter algorithm with the mapping of BP neural network, the nonlinear mapping of the BP neural network was used to split and select the weights. With the observed value regarded as the target signal of the neural network, the weights of small weight particles were increased by multiple training in the neural network, thus improving the diversity of particle filter algorithm. The simulation and experimental results showed that the BP neural network particle filter method could effectively filter out the noise signal in the package weighing system and improve the signal quality of the sensor. The proposed filtering method greatly improves the stability of the weighing system and effectively improves the accuracy of package weighing, and the said control method can significantly improve the stability, accuracy and robustness of the quantitative weighing control process.