Implementing wake models in wind farm control can reduce the negative impact of wake effects from upstream turbines on downstream turbines. However, variations in turbine types and environmental conditions often lead to discrepancies between model predictions and actual wake. To address this issue, this study conducts observational experiments on turbine wakes in wind farms using UAV-based wind speed measurement technology. Through turbine wake measurements, an optimization method was proposed to improve the FLORIS model, which is widely used in wake control. Comparative evaluations of the original and optimized models showed that the optimized model aligned more closely with the measured wake data. By applying wake models to the entire wind farm, the optimized model provided more consistent estimates of wake deficit with the actual wake deficit across the farm. For a region significantly impacted by wake effects, under wind speed conditions of 5 m/s to 10 m/s, the accuracy of the wake deficit estimated by the optimized model improved by 0.2% to 7% compared to the original model. Across all wind speed ranges, the precision of wake deficit estimation for the entire wind farm increased by approximately 1.5%. When applied optimized model to wake control within the wind farm, the potential to enhance annual energy production by approximately 0.22%.