This paper develops a multi-objective optimization approach for incorporating the conditional probability of fire flow failure into the design of branched water networks. To this end, a new analytical probabilistic model was developed to quantify the conditional probability of fire flow failure in branched networks and incorporated into the non-dominated sorting genetic algorithm (NSGA-II). The optimization sought to minimize capital cost through pipe diameter and pump selection and to minimize the conditional probability of fire flow failure. The NSGA-II was applied to two branched networks to generate Pareto-optimal solutions. Results indicated a strategic allocation of pipe and pump capacity with limited fiscal resources and with a reduction in uncertainty of fire flow failure. Interestingly, optimization results for a real branched network supported the industry practice of using a minimum 150 mm distribution main sizing to provide fire flow protection.