In this study, a new multiscale framework based on 2.5D discrete dislocation dynamics is proposed to predict the subsurface damaged layer evolution in micro-milling titanium alloy Ti-6Al-4V. This model takes into account the characteristics of component microstructure transformation and grain refinement by tracking the movement of matrix defects such as multiplication, slip, climb, cross-slip, junction and annihilation. Meanwhile, to understand the size dependence effect in micro-machining operation, a novel finite element orthogonal cutting model with dislocation density-based strain gradient constitutive equation is proposed and applied to reveal the far-field solution of driving stress of subsurface defect. The subsurface damaged layer characteristics including dislocation distribution and microstructure alteration of Ti-6Al-4V under various cutting conditions are studied with qualitative and quantitative assessment. The effects of processing parameters on subsurface features are analyzed. The obtained results have been compared with experimental findings utilizing the X-ray diffraction tests for validation purpose.