Abstract: In order to scientifically control monocotyledon weeds, explore the damage law of monocotyledon weeds on millet. Using the plot experiment and the method of SPSS 18.0 nonlinear regression analysis, to simulate and compare the main experience models. The results showed that, it was very significant positive correlation between millet yield loss and monocotyledon weeds density, power hyperbolic function model goodness-of-fit was maximum, as high as 99.6%, the sum of squared residuals (23.947) was minimum, it was the optimal model to simulate the competition relation between monocotyledon weeds and millet. Prediction equation was Y=d0.083/(0.793+0.013d0.083), the biggest competitiveness of interspecific and intraspecific in monocotyledon weeds were 1.2610 and 0.0164 respectively, millet yield maximum loss rate 76.92%, elastic coefficient r=0.083<1, millet yield loss increased slowdown with the increase of monocotyledon weeds density. The study provided theoretic support for comprehensive control monocotyledon weeds.