The maintenance planning of deteriorating bridges is to find a balance between obtained performance and incurred cost. Because the planning horizon spans tens of years, a certain amount of uncertainty is inherent in forecasting the deteriorating process and the costs and effects of maintenance actions. This paper proposes a multiobjective simulation optimization framework to establish the trade-off among the expected values of life-cycle maintenance cost and of the performance measures. The trade-off information, represented as the Pareto front, gives planners sufficient flexibility to respond to various needs. The optimization is performed by a multiobjective particle swarm optimization (MOPSO) algorithm, while Monte Carlo simulation is used to model the uncertainties. To alleviate the computational burden, the proposed framework is implemented in a parallel computing platform, where three programming paradigms (master-slave, island, and diffusion) are developed to distribute computation across processors and to control interprocessor communication. The validity of the proposed framework, along with the parallel paradigms, is investigated through a practical case. It is shown statistically that the proposed MOPSO algorithm is superior to the well-known nondominating sorting genetic algorithm II as the former can obtain a better Pareto front, whose convergence and diversity are measured together by the hypervolume indicator. Both the island and diffusion paradigms, being loosely synchronous, exhibit high efficiency and good scalability as they achieve superlinear speedups. The island paradigm outperforms the other two in terms of improved solution quality within fixed time.