The microstructure of machined surfaces is crucial to the overall performance and service life of mechanical structures. In this study, we performed cutting experiment, test analyze, and theoretical calculation to investigate the evolution in gradient microstructures of the superficial layer on machined surface, including grain size, dislocation density, grain boundary and misorientation angle, in the turning of high strength alloy steel with a self-developed coated carbide microstructure turning tool under different cutting velocities. Single-factor cutting experiments and corresponding tests were carried out using X-ray diffraction (XRD) and electron back scattered diffraction (EBSD) to investigate effects of cutting velocity on the microstructure of the machined surface during the cutting process. In addition, the performance and quality of the machined surfaces were assessed using a micro-hardness tester, laser scanning confocal microscopy, and atomic force microscopy. The relationship between microstructure and performance of the machined surface was explored by establishing a cutting model. The results showed that higher cutting velocities before exceeding the transition threshold with a larger relative evaluation coefficient can raise the thickness of the strengthening layer. This changes the gradient-distribution of dislocation density, grain boundary, and misorientation angle in the machined surface resulting in finer grains, thereby strengthening the machined surface. Moreover, the thickest agglomerate layer of equiaxed crystals was also observed under the higher cutting velocity, which was found to improve the performance of the workpiece (Micro-hardness of 318.07 HV and roughness of 0.05 μm), suggesting the approach to control cutting conditions to achieve better machined surfaces by evaluating the comprehensive function of mechanical load and thermal load in the metal machining process is effective.