Due to the destructive environmental effects of oily wastewaters and the required purification of these effluents, novel polycarbonate (PC) based mixed matrix membranes (MMMs) including two various kinds of halloysite nanotubes (HNTs) were fabricated by phase inversion technique. The additive utilized in the preparation of neat membrane was polyvinyl pyrrolidone (PVP), while sodium dodecyl sulfate (SDS) was used for the modification of HNTs. Both SDS-modified and unmodified HNTs were separately impregnated into the membrane matrix with different nanoparticle loadings and the obtained mixed matrix membranes were utilized in the ultrafiltration process of oil/water emulsion. The effects of additive and nanoparticle loadings were evaluated on some properties of the membranes, namely hydrophilicity, functional groups, surface morphology, and mechanical characteristic via contact angle measurement, Fourier transform infrared spectrophotometry (FTIR), field emission scanning electron microscope (FESEM), transmission electron microscopy (TEM), and mechanical strength tests, respectively. Based on the obtained results, the membranes containing SDS-modified nanofillers exhibited more hydrophilicity and mechanical strength than the mixed matrix membranes incorporated with unmodified HNTs. Compared to bare polycarbonate membranes, adding SDS-modified HNTs into the membrane casting solution improved the pure water flux of the synthesized membranes and reduced the filtration time for the acquisition of the complete oil removal yield (100%). As an example, the water flux of the PC-0.75MH MMM was enhanced around 319% in comparison with that of the bare polycarbonate membrane. Evaluation of the membranes antifouling characteristic displayed that the optimized PC-0.75MH membrane had a flux recovery of above 95% after 3 regeneration cycles. Furthermore, the water permeability was analyzed using D-Optimal design response surface methodology. The results of the modeling study showed that the experimental data had about 97.37% proximity with the proposed model.