Accurate measurement of three-dimensional green biomass is essential for assessing urban greening levels and conducting quantitative ecological benefit studies. Current remote sensing-based green biomass measurement methods require time-consuming field surveys to collect tree species data and establish species-specific crown morphological parameter models. This study develops an efficient green biomass estimation approach by simulating individual-tree green biomass using directly quantifiable parameters from high-resolution remote sensing imagery. First, centimeter-level UAV imagery was acquired across Fuzhou's urban areas, and remote sensing methods were employed to measure the individual and total green biomass in different areas (streets, neighborhoods, and urban areas). Next, the green biomass characteristics of different areas were compared and analyzed. Finally, regression models were constructed using directly quantifiable factors from remote sensing imagery to estimate the green biomass in various regions. Key findings reveal: (1) Street trees in Fuzhou show low species diversity (61.3% Ficus spp.), while neighborhoods maintain balanced species distributions. Significant green biomass variations exist across spatial scales (streets: 351.6 m³; neighborhoods: 143.7 m³; urban areas: 161.4 m³). (2) Regression models using canopy projection area and perimeter achieved high precision (adjusted R²: 0.921 for streets, 0.873 for neighborhoods, 0.882 for urban areas), demonstrating an accurate, efficient solution for multi-scale green biomass assessment.