<p>Surface modification technology significantly improves the friction performance of titanium alloy implants. However, traditional surface parameters are difficult to characterize the friction performance of coatings. Therefore, conducting an in-depth analysis of surface image features is of great significance for optimizing coating design and predicting friction performance. In study, Cu and hydroxyapatite (Cu-HAp) coatings are prepared on Ti6Al7Nb titanium alloy substrate using the jet electrodeposition technology, and surface roughness and average friction coefficient of coatings are measured. Surface images are acquired by machine vision system. Following grayscale conversion, image segmentation, and noise reduction, five image feature parameters are extracted. Besides, the most significant factor affecting the average friction coefficient of coatings is determined by the mutual information method. Finally, the Bayesian regression model combining with gradient-boosting decision tree is used to predict the average friction coefficient of the coating. The results reveal that image feature parameters have a greater impact on the average friction coefficient than surface roughness, thickness, and jet electrodeposition parameters. Among them, the inverse difference moment has the greatest influence, followed by entropy, contrast, and correlation. The accuracy of the model is 77%, with a mean square error of 0.0013 and a root mean square error of 0.037.</p>