The primary aim of the current study is to investigate the influence of input parameters of near dry electric discharge machine (ND-EDM) upon the output performances including the MRR, EWR, SR and WLT for the fabricated new metal matrix composite (MMCs) of aluminum A7075 matrix nanocomposites by adding 8% of Microscopic Slide Glass Nanoparticles (MSGNPs) as reinforcements to improve the metallurgical and mechanical properties of Al-7075/MSGNP composites using stir-casting method. In ND-EDM the dielectric medium plays a significant role in the procedure responses. In the current work, the vegetable oil with gases, such as air, Ar, mix (Ar+N 2 ), and Freon were used as a dielectric media. The obtained results show that the highest MRR achieved when using the vegetable oil + Freon gas, reached 29.425 mm 3 /min, and then 26.943 mm 3 /min when using the vegetable oil + Air as a dielectric. The lowest EWR achieved when employing the vegetable oil + Argon gas, reached 0.120 mm 3 /min, and then 0.175 mm 3 /min. The lowest SR values obtained for all the designed experiments reached 3.287 µm when using Ip (10 A), Ton (1600 µsec), and Ar additive gas, followed by 4.567 µm when adding Freon gases to the dielectric. In the ND-EDM, the average of recast white layer thickness in the case of vegetable oil + air, vegetable oil + Ar, vegetable oil + mix (Ar-N 2 ), and vegetable oil + Freon was 1.505, 1.180, 0.456, and 0 μm, respectively. These unique results can be used to increase the service and fatigue life of parts and machines that are exposed to sudden dynamic mechanical or thermal loads, without the need for additional operations to remove this brittle layer, which causes the failure of these parts with a short service life. The created mathematical models displayed a higher value of R-Square and the adjusted R-square, which manifest a better fit. Normal probability plots of the residuals for MRR, EWR, and SR elucidated an obvious pattern (i.e., the points were stabilized in a straight line) which indicates that every factor affects the mentioned responses and the outcomes of these responses from the regression model (predicted value by factorial) and the true values (from the experiments).