Occurrence of pharmaceutical micropollutants in aquatic environments has been one amongst serious environmental problems. During this study, two reactors, including a sequencing batch reactor (SBR) + powdered composite adsorbent (CA) (first reactor, SBR + CA) and a sequencing batch reactor (second reactor, SBR), were designed to treat synthetic wastewater. Powdered CA was added with a dosage of 4.8 g L−1 to the first reactor. Tap water was contaminated with chemical oxygen demand (COD), ammonia and three pharmaceuticals, namely, atenolol (ATN), ciprofloxacin (CIP) and diazepam (DIA) to produce synthetic wastewater. The SBR + CA illustrated a better performance during synthetic municipal wastewater treatment. Up to 138.6 mg L−1 (92.4%) of COD and up to 114.2 mg L−1 (95.2%) of ammonia were removed by the first reactor. Moreover, optimisation of pharmaceuticals removal was conducted through response surface methodology (RSM) and artificial neural network (ANN). Based on the RSM, the best elimination of ATN (90.2%, 2.26 mg L−1), CIP (94.0%, 2.35 mg L−1) and DIA (95.5%, 2.39 mg L−1) was detected at the optimum initial concentration of MPs (2.51 mg L−1) and the contact time (15.8 h). In addition, ANN represented a high R2 value (>0.99) and a rational mean squared error (<1.0) during the optimisation of micropollutants removal by both reactors. Moreover, adsorption isotherm study showed that the Freundlich isotherm could justify the abatement of micropollutants by using CA better than the Langmuir isotherm.