Abstract: To investigate the material removal mechanism of single-crystal silicon during ultra-precision cutting with a natural single-crystal diamond tool, experimental testing was combined with molecular dynamics simulations to systematically analyze the effects of tool rake angle and spindle speed on cutting performance and surface quality. In the experimental procedure, ultra-precision turning and surface characterization techniques were employed to measure the cutting force, friction coefficient, and surface roughness under different rake angles and spindle speeds. In the MD simulation, an atomistic cutting model was established to reveal the influence of cutting speed on atomic phase transitions and defect formation. It is indicated that a 0° rake angle tool significantly reduces average tangential force and friction coefficient, achieving superior surface finish. Increased spindle speed effectively decreases surface roughness and improves surface quality. Simulation results further demonstrate that higher cutting speeds suppress high-pressure phase formation and reduce defect atoms. Predictive models were constructed based on experimental and simulation data using Gaussian Process Regression and Bootstrap resampling for performance trend analysis and uncertainty assessment. Theoretical guidance and data support for optimizing process parameters and controlling surface quality in ultra-precision machining of single-crystal silicon are provided.