This paper aims to present a mathematical model for the design of a hybrid power system (HPS) to support a remote island with 100 thousand citizens. The goal is to reduce diesel fuel consumption by adequate expansion of wind power supply. Pumped hydroelectric storage (PHS) is used in the HPS to buffer the impact of intermittent behavior of wind energy. A superstructure is proposed for HPS design, considering all possible capital decisions (e.g. the number of wind turbines) and hourly-basis operational variables (such as the amount of surplus electricity in storage and its discharge). The HPS design problem can then be formulated as a mixed-integer linear program (MILP) based on the proposed superstructure. For a given total share of wind power, the optimal mix of diesel-based and wind power supplies as well as the required capacity of PHS are determined using a four-step optimization approach, involving minimizing (i) the consumption of diesel fuel, (ii) the number of wind turbines, (iii) the size of the upper water reservoir, and (iv) the charge/discharge rates of the PHS system. In this sequential optimization, the objective value obtained in a previous step is added as an additional constraint to the next step. The proposed HPS design model is applied to a real case study of the remote K Island on the other side of Taiwan Strait using hourly-basis, year-round historical data. Inclusion of other renewable energy sources, such as photovoltaic cells and biomass-fired power plants, as well as economic perspectives will be considered in future work. (C) 2015 Elsevier Ltd. All rights reserved.