Microstructured surfaces with subsurface fine crystallization have excellent mechanical properties. This study proposes a cooperative servo indenting (CSI) approach for accurately machining fine-crystallized microstructured surfaces on titanium alloys. Slow servo indentation is used in CSI to generate desired microstructured surfaces, in which large plastic strain is generated at the deep subsurface, resulting in subsurface dislocation concentration and grain refinement. Meanwhile, another fast servo motion (FSM) system cooperatively moves to compensate the springback errors according to the instantaneous indenting depth. To determine the compensating motions, a springback error model is proposed by considering the dynamic recrystallization (DRX), dislocation increment, indenting forces and subsurface strain-stress distribution. The proposed CSI approach and model are experimentally validated by machining different microstructured surfaces, such as hierarchical micro-grooves, micro QR codes, inverted micro-pyramid arrays, on Ti6Al4V alloys. The results show that springback errors lead to near 30% deviation between the machined microstructured surfaces and the desired ones, and the proposed CSI can efficiently reduce the springback errors from 4.26 to 0.62 µm. Through EBSD and nano-indentation analysis, it is also validated that highly increased dislocation density and DRX at the subsurface promote an increase of hardness of the microstructured surfaces. Besides, the grain size obviously reduces from 15 to 3 µm in the subsurface region at a depth of nearly 45 µm.