Abstract: Forest biomass is an important index to reflect the structure and functional characteristics of the ecosystem, and it is also the basis to study the energy flow and material circulation in the forest ecosystem. Under the condition of according with the accuracy and reliability, it is necessary to improve sampling technique and make use of historical data to reduce the sampling units and make them concentrated. Spatial distribution pattern of biomass is an important factor affecting the accuracy of biomass inventory, and sampling technique is a key factor affecting the efficiency of the survey. The number of sample units in traditional equal probability sampling is large and scattered under the set accuracy and reliability, which has limitations in biomass inventory. The research on adaptive sampling technique can carry out unequal probability sampling based on the spatial distribution pattern of forest biomass, effectively improving the investigation efficiency, so as to meet the actual needs of biomass estimation at different regional scales and spatial distribution characteristics.