Wind farms with clustered wind turbines are the primary form for capturing and utilizing wind energy, and the rapid and accurate prediction of wind farm power generation is crucial to ensuring the efficient operation of wind power projects. To address the insufficient accuracy of traditional power generation assessment methods, which stems from their failure to fully account for the complex flow field environment of wind farms, incoming wind shear effects, and inter-turbine wake effects, this study proposes a novel refined power generation assessment method based on a three-dimensional wake model considering atmospheric stability and a novel equivalent wind speed calculation model. Three types of wind farms were selected as research objects, including two conventional wind farms, wind farms with vertical staggered layout, and wind farms under inflow conditions with different atmospheric stabilities, with their power generation data used to systematically verify the proposed method’s accuracy and generality. The results indicate that for conventional wind farms, the novel power generation assessment method exhibits superior performance compared with traditional methods, and achieves prediction accuracy comparable to Large Eddy Simulation (LES) under various wind conditions. For wind farms with vertically staggered layouts, the proposed method can more accurately evaluate the wake losses of downstream turbines that are only partially located in the wake region of upstream turbines, enabling refined assessment of the power generation of each individual turbine with good agreement with LES results. Furthermore, the proposed method demonstrates significant advantages in power generation prediction under different atmospheric stabilities. Due to the integration of the wake model considering atmospheric stability, the assessment accuracy of the proposed method is significantly higher than that of traditional methods under both stable and unstable atmospheric conditions. Overall, the novel assessment method comprehensively accounts for the coupled influence mechanism of atmospheric stability, wind shear, and wake effects on the complex flow field within the wind farm, thereby enabling more accurate prediction of the actual power generation of each wind turbine.