Introduction: The participation of conductive powders in Electro-Discharge Machining (EDM) holds a great potential to improve the machining process. In this study, tungsten powder alloy added to the dielectric liquid during the EDM process, is called Powder Mixed EDM (PMEDM) to process heat-treated SKD61 steel was investigated. The aim of this study comprises: (i) considering the influence of essential process parameters, embracing pulse-on time (Ton), peak-current (Ip), and amount of powder (Ap) on tool wear rate (TWR) and material removal rate (MRR), and (ii) finding an optimation coalescence of the process variables for enhanced the MRR, and the reduced TWR. Methods: For this aim, the Box-Behnken matrix was adopted for the experiment design, and a series of 15 experiments has been performed to obtain empirical data. Subsequently, the adequate mathematical models for MRR and TWR were instituted, and the analysis of variance (ANOVA) was applied to assess these models' adequacy. Finally, grey relational analysis (GRA) was adopted for the multi-attribute optimization. Results: The results revealed that Ip proves the most robust influence on MRR and TWR. However, the proceeding influence is Ton and Ap for MRR, while this reverse is for TWR. The predictive models of MRR and TWR were constituted and validated with the adequacy/precision through coefficients (comprising “R2” of MRR and TWR corresponding to 0.9899 and 0.9918 , “R2(pred)” of MRR and TWR corresponding to 0.8504 and 0.8699, and “R2(adj)” of MRR and TWR corresponding to 0.9716 and 0.977). From the predictive models, the optimal responses and process variables, including MRRmax of 0.003397818(g/min), TWRmin of 0.000481408(g/min), peak-current of 5(A), pulse-on time of 150(µs), and powder concentration of 15(g/l) were found. In addition, the comparison of the micro-defects at the optimum electrical mode was conducted between having powder mode and having no-powder mode was conducted. As results, the surface obtained with the powder mode has fewer micro-cracks, voids, droplets, and smaller globules of debris than that of the surface obtained with the powderless mode. Conclusions: The results attained from evaluation of the influence of process parameters on the machining performances, establishment of a prediction model for the machining performances, and optimization of process parameters, which can be applied in factual mold manufacturing, and helps technologists and researchers having the most suitable choices. Besides, the methods applied in this study can be applied in the PMEDM process to study different powders and workpiece materials.