This paper presents an improved optimization method for sizing standalone photo-voltaic (PV) systems. Numerical methods for optimization and configuration of array size, charge controller, storage battery capacity, and power inverter of PVs are implemented using MATLAB and hourly meteorological data and load demand. DC voltage operation is also optimally selected. Loss of load probability analysis is conducted to set the benchmark for determining all possible PV arrays and battery capacities. Then, the optimum design is proposed based on the lowest levelized cost of energy and net present value. The case study refers to the meteorological data of Yemen and the typical load role of the domestic sector in Yemen. Results show that the net present value of 6.6024 kWh/day PV system for Yemen is 22224 USD, while the cost of energy generated by the proposed system is 0.403 USD/kWh and the loss of load probability (LLP) is equal to 0.130%. (C) 2018 Elsevier Ltd. All rights reserved.