Significant advancements in ultra-precision machining technology have forced a re-examination of the spindle perpendicularity errors' impact on the milled surface quality at the micro-nano scale. In this paper, a method of spindle precision adjustment is proposed to enhance surface finish quality. Sensitive errors in the machining process are identified using multi-body kinematic theory, with the milling process serving as an example. A two-degree-of-freedom (2-DOF) rotation platform is designed, optimized, and fabricated. The platform's static model is established based on elastic beam theory and verified by finite element analysis. Structural parameters are optimized via the response surface method in combination with the Pareto front. Experimental results reveal the effects of spindle speed, voltage amplitude, vibration frequency, cutting depth, and feed rate on the platform's modulation performance. The static modulation experiment shows that the perpendicularity error between the spindle and the guideway can be reduced from 92.5 μrad to 0.25 μrad. Finally, milling experiments show that the surface quality can be improved by 37.6% after spindle modulation.