An improved method to determine cutting force coefficients for bull-nose cutters is proposed based on the semi-mechanistic cutting force model. Due to variations of cutting speed along the tool axis in bull-nose milling, they affect coefficients significantly and may bring remarkable discrepancies in the prediction of cutting forces. Firstly, the bull-nose cutter is regarded as a finite number of axial discs piled up along the tool axis, and the rigid cutting force model is exerted. Then through discretization along cutting edges, the cutting force related to each element is recalculated, which equals to differential force value between the current and previous elements. In addition, coefficient identification adopts the cubic polynomial fitting method with the slice elevation as its horizontal axis. By calculating relations of cutting speed and cutting depth, the influences of speed variations on cutting force can be derived. Thereby, several tests are conducted to calibrate the coefficients using the improved method, which are applied to later force predictions. Eventually, experimental evaluations are discussed to verify the effectiveness. Compared to the conventional method, the results are more accurate and show satisfactory consistency with the simulations. For further applications, the method is instructive to predict the cutting forces in bull-nose milling with lead or tilt angles and can be extended to the selection of cutting parameters. An improved method to determine cutting force coefficients for bull-nose cutters is proposed based on the semi-mechanistic cutting force model. Due to variations of cutting speed along the tool axis in bull-nose milling, they affect factors significantly and may bring remarkable discrepancies Firstly, the bull-nose cutter is regarded as a finite number of axial discs piled up along the tool axis, and the rigid cutting force model is exerted. Then through discretization along cutting edges, the cutting force related to each element is recalculated, which equals to differential force value between the current and previous elements. influences of speed variations on cutting force can be derived Compared to the conventional method, the results are more accurate and show satisfactory consistency with the simulations. For further applications, the method is instructive to predict the cutting forces in bull-nose milling with lead or tilt angles and can be extended to the selection of cutting parameters.