Hydroxyl radical (•OH) scavenging demand can be an indicator that represents the water quality characteristics of raw water. It is one of the key parameters predicting UV/H2O2 system performance and affects the operating parameters. Based on the •OH scavenging demand, we developed a model predictive control strategy to meet the target compound removal efficiency and energy consumption simultaneously. Selected pharmaceutically active compounds (PhACs) were classified into three groups depending on the UV direct photolysis and susceptibility to •OH. Group 1 for photo-susceptible PhACs (acetaminophen, amoxicillin, diclofenac, iopromide, ketoprofen, and sulfamethoxazole); group 2 for PhACs susceptible to both direct photolysis and •OH oxidation (bisphenol A, carbamazepine, ibuprofen, naproxen, ciprofloxacin, and tetracycline); and group 3 for photo-resistant PhACs (atenolol, atrazine, caffeine, and nitrobenzene). The results of modeling to achieve 90% removal of PhACs at N and B plants were as follows. For group 2, the optimized operating parameter ranges were as follow (N plant: UV 510–702 mJ cm−2, H2O2 2.96–3.80 mg L−1, EED 1088–1302 kWh m−3; B plant: UV dose 1179–1397 mJ cm−2, H2O2 dose 3.56–7.44 mg L−1, EED 1712–2085 kWh m−3). It was confirmed that the optimal operating conditions and EED values changed according to the •OH scavenging demand.