| Turning process is a machining process that is often found in the automotive industry. One of the factors that determine the quality of the turning process is the surface roughness of the workpiece. Experiments were conducted to test how much influence the turning process variables against the workpiece surface roughness, created a model of surface roughness by using neural networks and multiple linear regression method, and compared the errors produced between the two models of surface roughness. This study used four independent variables consisting of cutting speed, feed, depth of cut and hardness of the workpiece. Workpiece used were carbon steel shaft-type DIN ST 37and AISI 1045. Turning tool used was carbide tool. Surface roughness model formulation used neural networks and multiple linear regression. Based on the experimental results, feed has the greatest influence on the value of the surface roughness. Depth and cutting speed are in the second and third positions. Hardness of workpiece does not have a significant influence on the surface roughness. Has successfully formulated a numerical model to predict surface roughness by using neural networks and multiple linear regression method |