Ultrafiltration membranes have several advantages over conventional drinking-water treatment. However, this technology presents major limitations, such as irreversible fouling and low removal of natural organic matter. Fouling depends heavily on the raw-water quality as well as on the operating conditions of the process, including flux, permeate recovery, pre-treatment, chemical cleaning, and backwashing. Starting with the premise that the optimisation of operating variables can improve membrane performance, different experiments were conducted in a pilot plant located in Granada (Spain). Several combinations of permeate and backwashing flow rates, backwashing frequencies, and aeration flow rates were tested for low-quality water coming from Genil River with the following results: the effluent quality did not depend on the combination of operating conditions chosen; and the membrane was effective for the removal of microorganisms, turbidity and suspended solids but the yields for the removal of dissolved organic carbon were extremely low. In addition, the threshold transmembrane pressure (-0.7 bar) was reached within a few hours and it was difficult to recover due to the low efficiency of the chemical cleanings. Moreover, greater transmembrane pressure due to fouling also increased the energy consumption, and it was not possible to lower it without compromising the permeate recovery. Finally, the intensification of aeration contributed positively to lengthening the operation times but again raised energy consumption. In light of these findings, the feasibility of ultrafiltration as a single treatment is questioned for low-quality influents.