In the production of refined sunflower oil, waxes are removed during the winterization stage, and wax crystals are separated through filtration assisted by filtration aids. Commonly used filtration aids in oil refining include perlite and diatomaceous earth. After winterization, a significant amount of filter cake remains as a by-product and is treated as waste. Today, natural cellulose fibers are being promoted as filtration aids. Their advantages are numerous, both in the production process and from an environmental perspective. However, their only disadvantage is their higher cost. Therefore, in this study, 57 filtration cycles during the industrial sunflower oil winterization step using cellulose-based filtration aids were monitored. Different process parameters, including the pressure differential on the filter, the flow rate of filtered oil, constant pressure period, the quantity of filtered oil, filtration time, the quantity of pre-coating and dosing filtration aids, the volume of filtered oil, the concentration of dosing filtration aid, as well as the mass of separated waxes, were observed. Additionally, artificial neural networks were applied to predict process parameters, optimize the process, and, above all, determine the dosage of filtration aids, which will make the process more economical. The optimal filtration process is performed at a pressure differential of 3.3 bar, lasting a total of 39 h, with 32 h at constant pressure, resulting in 322,503 kg of filtered oil and 90.41 kg of waxes. The optimal quantity of cellulose-based filtration aids employed for pre-coat was 80 kg, and for dosing, 375 kg, with an optimal concentration of 0.12% w/w.