As for optimizing performance of chlorine and trihalomethane (THM) in water distribution systems, an expression measuring THM performance and a derivation of THM-I/O model were proposed. And a multi-objective optimization model considering maximizing level of chlorine performance, maximum level of THM performance and minimizing cost of chlorination was also established. Four factors, including bulk chlorine decay coefficient, coefficient of proportionality between THM and chlorine consumption, initial THM concentration in treated water and number of boosters, were evaluated to study their effects on Pareto front of the multi-objective optimization model. Chlorine decay and THM generation were simulated using EPANET_MSX. An improved non-dominated sorted genetic algorithm-Ⅱwith integer code was proposed to optimize the model, and Pareto fronts could be obtained under different conditions. Results show that with the decrease of bulk chlorine decay coefficient, cost of chlorination can be reduced and level of THM performance will be improved. Diminishing coefficient of proportionality between THM and chlorine consumption and initial THM concentration out of reservoir can both improve the level of THM performance, but can not affect the cost of chlorination. Increasing number of boosters can reduce dosage of chlorine, but booster station construction will increase total cost of chlorination. There-fore, reducing the bulk chlorine decay coefficient is the most economical and effective method.