The presence of free chlorine ions in industrial wastewater poses significant environmental and operational risks due to their corrosive, oxidative, and toxic nature. Among various dechlorination techniques, activated carbon filtration has proven effective due to its high surface area and strong adsorption capacity. However, optimizing the amount of activated carbon used while ensuring maximum chlorine removal remains a practical challenge. This study presents a MATLAB-based multi-objective optimization approach for modeling and minimizing activated carbon dosage while maximizing dechlorination efficiency. A laboratory-scale experimental setup was developed to simulate the filtration process. A total of 200 experiments were conducted by varying input parameters such as flow rate (10–100 m³/h), initial chlorine concentration (1–10 mg/L), pressure (1.5–5 bar), pH (6.5–8.5), temperature (15–35°C), and activated carbon dose (50–200 kg). These parameters and corresponding residual chlorine concentrations were used to train a neural network in MATLAB using the Levenberg–Marquardt backpropagation algorithm. The network achieved high predictive accuracy with an MSE of 0.00176 and R² = 0.9912. A built-in optimization function in MATLAB was then used to identify the optimal combination of input variables that minimized chlorine levels with the least amount of activated carbon. Results showed that a dose of 84 kg of activated carbon reduced residual chlorine to 0.02 mg/L, maintaining a 98% removal efficiency. This research demonstrates the potential of combining experimental data with intelligent modeling techniques to support sustainable and cost-effective wastewater treatment solutions in industrial applications.