The spatial distribution pattern of a population demonstrates a combination of the unique characteristics of the population, and the interrelationship between species and the environment. In the present study, Stipa breviflora was used to examine the point pattern and spatial correlation of primary plant populations using Programita software, Ripley's K function, and the Monte Carlo Method. The results showed that the scale of random distribution gradually increased with elevation of grazing intensity, or continuous grazing, and may be the primary explanation for the clustered distribution of populations at large scales. The clustered distribution scale of S. breviflora increased in the "grazing prohibited" and "overgrazed conditions". These two conditions were also associated with a convergence trend in S. breviflora. Increasing grazing intensity, or continuous grazing, gradually amplified the random clustered distribution scale of S. breviflora and may be the primary explanation for this clustered distribution. On a larger scale, Cleistogenes songorica showed a clustered distribution under "heavy grazing in spring+heavy grazing in summer+light grazing in autumn", and "heavy annual grazing". The critical point of the above scale conversion showed an increasing trend with grazing. Allium polyrhizum showed a trend of clustered distribution in response to grazing disturbance. The clustered distribution of this species was reduced to improve the population stability and avoid interference. Plant populations had specific response strategies under different grazing intensities, including enhancement of species sociality, and reduction of the clustered distribution, to improve population stability and avoid excessive disturbance. S. breviflora, C. songorica, and A. polyrhizum had either no correlation or a negative correlation at the small scale. At larger scales, primary plant species had either no correlation or a positive correlation. At a scale of 0-250 cm,S. breviflora and A. polyrhizum showed no correlation in conditions of "heavy grazing in spring+heavy grazing in summer+light grazing in autumn". However, at larger scales, S. breviflora and A. polyrhizum were positively correlated in the conditions of "zero grazing in spring+heavy grazing in summer+light grazing in autumn", and "zero grazing in spring+light grazing in summer+heavy grazing in autumn". Conditions of "spring grazing", "summer grazing", and "fall grazing" produced relatively lager-scale population patches that benefited the stability of the population. Thus, these methods are more suitable for the desert grassland.