The real estate property maintenance problem (REPMP) deals with the problem of the maintenance of real estate property buildings' components using a limited budget in a limited period to achieve multiple and often conflicting objectives. Maintenance of the components requires a multiannual maintenance action plan optimized simultaneously on different criteria relative to buildings' maintenance objectives (the quality of service, client satisfaction, regulatory compliance, etc.) while respecting a predetermined budget. This article presents an innovative modeling of REPMP into a multiobjective multidimensional knapsack problem and its optimization using a generalization of the discrete multiobjective particle swarm optimization algorithm and the Pareto optimality. An application of this methodology using data from a real estate property of 22 buildings of a major French company is presented and discussed.