Although there have been numerous attempts to define how different laser polishing parameters affect the generated surface roughness, there has been no detailed investigation of how their effects can be combined to optimize the process. This paper applies statistical analysis to model and predict the resulting surface roughness for laser post-processing of components made of Ti-6Al-4V and produced by laser powder bed fusion. This model is based on analysis of a wideranging experimental programme investigating how the interaction of the governing parameters, i.e., laser power, number of repetitions, axial feed rate, scanning speed, and focal position, affected surface roughness. The experimental programme was the result of a robust Design of Experiments analysis and experimental analysis using ANOVA. It is expected that the outcomes will contribute towards the understanding of how the governing parameters influence the laser polishing process and final surface roughness, and would be a tool for optimizing their selection. The results of the ANOVA (analysis of variance) revealed that the most significant parameters are scanning speed followed by laser power and then axial feed rate. In addition, a clear tendency for the estimated Ra to decrease with the increase in laser power at lower values of axial feed rate and of scanning speed, and a focal position in the region of 2 mm. It is noted that the process parameters were varied over wide ranges, including extreme values, which made it difficult to accurately model the dependent variable over the full range of experimental trials.