Electrical discharge machining (EDM) is a prominent non-traditional manufacturing process for machining hard and electrically conductive materials. This study examines the effects of three key parameters, which are pulse duration (Ton), pause time (Toff), and discharge current (I) on the surface roughness (Ra) of C45 steel. Experimental studies were conducted using a design of experiments (DOE) approach to explore the parameter space. Data were analyzed through statistical methods and machine learning (ML) techniques to elucidate complex parameter interactions and non-linear relationships. Results reveal that the interplay between factors and Ra is inherently non-linear, making traditional linear regression models inadequate for accurate prediction. This study underscores the necessity of employing ML algorithms, to predict the EDM outlines. These findings contribute to enhanced process control in EDM, emphasizing the importance of adopting data-driven methodologies for improving surface quality in precision machining applications.