Spatial autocorrelation analysis was a statistical method of spatial statistics, and it could reveal the spatial structure of regional variables.By using the method of spatial autocorrelation analysis,the global spatial autocorrelation and local spatial autocorrelation of the population distribution of Fuzhou city was studied according to the demographic data of the streets and the towns in Fuzhou area in 2004.The results showed that the global Moran's I index was 0.409 3,which achieved the significant positive correlation and indicated that the population distribution of Fuzhou took on significant spatial cluster on global level.Meanwhile,significant local spatial autocorrelation of population distribution also occurred.The towns/streets of High-High correlation were mainly those of Fuzhou's central area and Fuqing,while the towns/ streets of Low-Low correlation were those of areas located far from Fuzhou central district.Towns' streets with significant positive correlation of population distribution were more than those with significant negative correlation on the whole.