The work aims to improve the accuracy of dynamic quantitative weighing packaging and the accuracy of data acquisition and signal measurement. The composition and working principle of the quantitative weighing system were analyzed, and based on the noise in the system, a weighing packaging filtering algorithm based on wavelet packet filtering was proposed. A fast discrete wavelet packet transform was realized by the tower decomposition method, and the wavelet packet decomposition coefficients were obtained by the discrete convolution equation, and thus the reconstruction of the filtering algorithm was realized. The simulation and experimental results showed that the wavelet packet filtering method can effectively filter the noise in the dynamic weighing signal, and improve the quality of the useful signal. The filtering algorithm improves the stability of the dynamic quantitative weighing system and greatly enhances the weighing packaging accuracy.