Purpose To scientifically estimate the biomass of forest vegetation in Chengdu and explore its characteristics, so as to help achieve the goal of “Dual Carbon” in Chengdu.

Method Based on the data of Chengdu forest resources planning and design survey plot (2017), the biomass of the forest resources was calculated, and the vegetation index was extracted based on Landsat 8 OLI images in 2017. The remote sensing biomass models was constructed by using multiple stepwise regression model, random forest model, and BP neural network model respectively, and the optimal model was selected to retrieve the forest vegetation biomass of the whole city, and its spatiotemporal characteristics were analyzed. Result The results showed that: (1) Among the three models, the random forest model had the best estimation performance, with R² of 0.521 and RMSE of 25.20 t/hm². (2) The estimated value of forest vegetation biomass in Chengdu was 1.6470×10⁷ t, and the average forest vegetation biomass in Chengdu was 56.89 t/hm². (3) The forest vegetation biomass in Chengdu showed significant spatial heterogeneity, mainly concentrated in Longmen Mountains in the northwest and Longquan Mountains in the middle and east, but the forest vegetation biomass level in Longquan Mountain areas was lower than that in Longmen Mountain areas. Terrain gradient analysis showed that forest vegetation biomass present a unimodal distribution pattern, and the peak area was located in the middle altitude area and slope regions. In the aspect of the slope, the semi-sunny slope had the highest proportion of biomass, and the ratio of forest vegetation biomass on shaded and semi-shaded slopes to that on sunny and semi-sunny slope was close to 1:1.

Conclusion The distribution of forest vegetation biomass in Chengdu had obvious significant spatial heterogeneity, showing a situation of more in the west and less in the east. It is necessary to adjust measures to local conditions, protect the ecological environment of Longmen Mountain area, and prioritize planting high carbon sequestration tree species in Longquan Mountain area to reasonably increase the forest vegetation biomass in Longquan Mountain area.