Abstract: Land use/land cover change (LUCC) simulation is an important content of land science. With the deepening of global change research, the LUCC simulation has gradually penetrated into crop area. The simulation of long-time spatial change of crops is helpful to reveal the complex relationship between "Human Nature" complex in the process of agricultural production. Currently, there is little research on crop spatial simulation. The existing models have technical limitations in spatial mapping and object recognition, and the application of models is mainly concentrated on grain crops. In this paper, builds a spatial production allocation model based on suitable probability to simulate the spatial distribution of citrus production in Sichuan Province. Results show that the model better reflects the statistical characteristics of citrus from 1980 to 2015, and the relative errors of statistical area and simulated area in each period are less than 25%. From the county level, the correlation coefficient between statistical area and simulated area is 0.987 6 ~ 0.999 9, showing a significant linear correlation. From 1980 to 2015, citrus area in Sichuan Province expanded rapidly, and the temporal change of citrus simulated area was consistent with the statistical area. Citrus space is mainly concentrated in the hilly area of central Sichuan and the plain area of Chengdu. The spatial pattern changes from scattered to regional agglomeration, roughly forming three citrus production concentration areas: Chengdu plain, southern Sichuan and northeast Sichuan. The error of the SPAM-Citrus model is affected by spatial resolution, spatial allocation rules, land use, crop distribution, and sampling points. In the future application of this simulation model, this paper suggests to select the sampling points with long planting years and select the appropriate resolution and spatial allocation rules for the research object and region to improve the accuracy of the prediction results.