The work aims to maximize the quantitative speed of particle packing scale. The main factors affecting the quantitative speed were analyzed from the perspective of reducing the quantitative period. Different feeding speed and the coarse and fine feed switching modes were set for numerous experiments by using DB8320 type granule packing machine. According to the test results, Matlab software was used to fit the nonlinear regression equation between the variables. The coarse and fine feeding amount and the feeding speed were calculated with the known parameters and the quantitative period formula. The best feeding mode was set and weighing was made again. Then measured data of other different modes were compared and analyzed. The results showed that quantitative average period was 5.48 s and the precision was 0.29% while weighing 10 kg rice with the above best methods. In conclusion, the quantitative speed was significantly improved under the premise of ensuring accuracy through experiments and optimum feeding mode by mathematical optimization.