Abstract: According to the fully observed data of 10.4 ha (400 m×260 m) large permanent plot of broadleaf-Korean pine forest established in Shengshan Nature Reserve, Heilongjiang Province in 2012, the moving window method and GIS were used to research the scale effect of spatial pattern of tree biomass and the minimum area of community spatial structure, and then computer simulation of the random sampling was used for verification. The results showed that there existed a minimum area of 0.49 ha in community structure of broadleaf-Korean pine forest in Shengshan Nature Reserve, and this area not only had the same biomass density as the entire plot, but also could reflect the spatial variability of biomass of the broadleaf-Korean pine forest in Shengshan. It accurately reflected the characteristics of community structure, providing a theoretical basis for more accurate estimation of biomass of the broadleaf-Korean pine forest. The minimum area determined by the moving window method was verified by the computer simulation of random sampling. For the sampling in the entire plot, only seven 20 m×20 m quadrats were needed to get the result with relative error less than 20% at 95.9% probability, and sampling in the minimum area also had the same effect. It is a good guidance for the investigation of the plot size of forest biomass in broadleaf-Korean pine forest. Based on the above research, we put forward the minimum area as an index to measure the complex degree of spatial structure of forest community. The minimum size of the forest community represents the range of spatial variability, i.e., the larger the minimum area is, the higher the spatial heterogeneityis.