Research has proven the adaptation of nitrite-oxidizing bacteria to unfavorable environmental conditions, and this work presents a novel concept to prevent nitrite oxidation during partial nitrification in wastewater. The approach is based on the real-time updating of mathematical models of the process to search for optimal set-points of pH and oxygen concentration in a continuous activated sludge reactor with a high sludge age (20.3 days). A heuristic optimization technique by 13 optimum set-points simultaneously maximized the degree of ammonia oxidation (α) and nitrite accumulation (β), achieving an (α + β) = 190% per day. The activated sludge reactor was conducted for 780 days under three control schemes: open-loop control, fuzzy model supervisory control and phenomenological supervisory control. The phenomenological supervisory control system achieved the best results, simultaneously reaching 95% ammonium oxidation and 90% nitrite accumulation. The Haldane kinetics were analyzed using steady-state concentrations of all nitrogen species, concluding that the simultaneous maximization of α + β led to selecting set-points at the extreme values of the following ranges: pH = 7.5–8.5 and DO = 0.8–1.0 mg O2/L, which enabled the inhibition of one nitrifier species. At the same time, the other one was relieved from inhibition. The 16sRNA assays indicated that the nitrite-oxidizing bacteria presence (genera Nitrobacter and Nitrospira) shifted from 32% to less than 8% after 280 days of continuous operation with optimal pH and oxygen set-points.