Modern manufacturing requires close monitoring by correlating finer frequency regimes of the surface texture to process parameters. This paper presents a study of B-spline wavelet-based multiresolution analysis (MRA) for surface texture characterization and process parameter correlation in end-milling. Motivated by multiresolution curves in computer graphics, an initial B-spline surface is first interpolated to the measured points of the surface texture. With a B-spline wavelet transform, the initial surface is decomposed into higher-frequency details and lower-frequency approximations. By taking the reconstructed surface roughness, waviness, and form based on different frequency regimes as the responses, regression models are created by considering the feed rate, spindle speed, axial depth of cut, and tool wear as controllable variables. The case studies and comparisons with an International Organization for Standardization (ISO) Gaussian filter have demonstrated the effectiveness of the proposed B-spline wavelet-based MRA algorithm for a surface texture analysis and manufacturing process diagnostics.