Abstract: Based on the spatial statistical analysis methods, the authors in this paper studies the spatial autocorrelation of particle size components in marine sediments using semi variogram function, Moran index, and geographically weighted regression. The surface sediments in the western North Yellow Sea were taken as the research object, and the main factors affecting particle size distribution were analyzed. Besides, the applicability of the main spatial statistical methods in analyzing marine sedimentary dynamics was evaluated. The results indicated that the range value indicated by the semi variogram function was consistent with the lag distance of the Moran index and had a good correspondence with the particle transport distance in the study area. The Moran's index anisotropy indicates that the study area had a main runoff direction of 135°, with a large lag distance, and relatively small lag distances in other directions, which was consistent with the main transport direction of the particle size trend model. The analysis of cold and hot spots showed that the northern coast of Zhifu Island is a sand content cold spot area and an clay content hot spot area, while the central part of the study area was an insignificant area. Geographically weighted regression model also showed that sediment particles didn't change linearly with water depth. Comprehensive analysis showed that the northern coast of Zhifu Island and areas with deep water depth of 40 m in the northeastern of the study area were mainly controlled by weak currents and cold water masses, respectively, While the central part of the study area was affected by the 135° direction current in Bohai Strait, resulting in an increase in sand content. This study could expand the scope of research on the distribution and influencing factors of marine surface sediments, and proposed some ideas for larger scale spatial heterogeneity studies.