<p>Tool-tip dynamics of the spindle are vital for predicting the milling stability thereby enhancing processing efficiency and precision in thin-walled workpiece milling. However, no reliable approach exists for accurately capturing and predicting tool-tip operational modes in response to complex and variable milling conditions, such as changes in spindle speed, spindle-holder-base position, and tool-holder combination. To address this issue, operational modal analysis is employed to identify tool-tip operational modes across various orders for different spindle speeds and spindle-holder-base positions. A mechanism-fused Gaussian process regression (GPR) model is then put forth to forecast operational modes that are dependent on speed and position for different tool-holder assemblies. The GPR model outperforms common regression models, showing a lower generalization error. Furthermore, the proposed mechanism-fused GPR model demonstrates an average prediction error below 16% for natural frequencies and under 13.8% for damping ratios. The established model is then applied to analyze and predict stability in thin-walled workpiece machining. A notable enhancement in stability prediction accuracy is achieved for milling operations conducted under the aforementioned complex and variable conditions when utilizing tool-tip operational modes derived from the mechanism-fused GPR model, compared to both experimental results and traditional stability prediction methods. In conclusion, the proposed mechanism-fused GPR model, which accounts for changes in spindle speed, spindle-holder-base position, and tool-holder combination, enhances the accuracy of stability prediction in complex and variable milling conditions.</p>