Separation chamber is an important part of the seed airscreening machine, and therefore the values of its fan frequency and air outlet magnitude have significant effects on the quality of seed separation. The RNG k-ε turbulence model and the DPM discrete phase model were adopted to simulate the separation process of seeds in the separation chamber. Then, the simulation results were analyzed by means of polynomial fitting, and the relationship between the working parameters, i.e. fan frequency and air outlet magnitude, and the cleaning performance, i.e. cleaning rate and loss rate were acquired. Based on this relationship, cleaning efficiency can be determined by calculating cleaning rate and losing rate, which synthetically reflects the quality of cleaning effect under different working conditions. It is convenient to analyze optimal working parameters of separation chamber by 3D map and contour map of the cleaning efficiency. By this means, the optimal working parameters of the fan frequency and air outlet magnitude for Chinese cabbage seeds, spinach seeds and pakchoi seeds, were obtained, which are 35Hz and 75%, 42Hz and 80%, 36Hz and 80%, respectively. The values of corresponding cleaning efficiency are 88%, 84% and 88%. The verification test of Chinese cabbage seeds in the real machine was designed, by comparing the cleaning performance of simulation with test, found that the results of simulation and test were consistent, which proves reliability of gassolid twophase flows model in separation chamber. The study provides reference for design and optimization of the separation chamber of seed airscreening machine.