Picea crassifolia,as a constructive species of forest ecology in Qilian Mountains,plays a very important ecological role in maintaining forest biodiversity and soil and water conservation. It is of great significance to explore the temporal and spatial dynamic characteristics of its natural regeneration spatial pattern for population regeneration. Fixed plots were established in P. crassifolia forests at altitudes of 2800 m,3000 m,and 3300 m,recoring the diameter at breast height (DBH),crown width,and coordinates of all P. crassifolia trees with DBH<50mm were recorded in 2017 and 2022. The influence of DBH of P. crassifolia on spatial distribution characteristics was explored by K function and univariate marker point pattern. Moran's I statistics and Zα statistics were used to analyze the spatial correlation characteristics of DBH of P. crassifolia. LISA clustering method was used to identify the spatial interaction relationship of DBH of each P. crassifolia. The results show that:(1) The interannual variation of the spatial pattern of new seedlings generally demonstrated an increasing trend in aggregation. The aggregation degree of 2800 m plot changed the most,while the aggregation degree of 3000 m and 3300 m plot changed little. In 2017,there was an alternate phenomenon of aggregation distribution and random distribution with the increase of spatial scale in 2800 m and 3000 m plot. By 2022,this phenomenon has largely disappeared. (2) The univariate marker pattern based on DBH could keenly capture the effect of DBH on spatial pattern,but this effect gradually disappeared with the increase of aggregation degree. (3) The interannual variation of the spatial correlation of DBH in the 2800 m and 3300 m plots showed a strengthening trend of positive spatial autocorrelation,showing a significant positive spatial autocorrelation in 2022,and the 3000 m plot showed a strengthening trend of negative spatial autocorrelation,showing an insignificant negative spatial autocorrelation in 2022.(4) The DBH agglomeration pattern among new seedlings at the 2800 m plot changed greatly,and the number of high-high and low-low agglomeration increased significantly. The DBH agglomeration pattern of the new seedlings in the 3000 m plot did not change much,and there were basically no plants with significant spatial correlation. To conclude,considering that the impact of the increase in the degree of aggregation is complex,including both negative and positive effects. It is suggested that the seedlings of P. crassifolia showing low-low agglomeration can be properly cut off in the future management,so as to reduce the competitive pressure and improve the survival rate of seedlings of P. crassifolia.