Conventional power generation is one of the main contributors to the phenomenon of the greenhouse effect. This has led to a diversification of electricity sources including environmentally friendly energy sources such as solar energy. Off-grid PV systems have gained some traction due to their cost-effectiveness for rural communities. However, the intermittent nature of solar is the main challenge to developing the off-grid PV system. Moreover, the high capital cost of PV systems as well as the storage batteries becomes the main concern for all PV users. Thus, this study aims to optimize the size of the PV system and battery simultaneously and design a cost-effective off-grid photovoltaic system considering various aspects such as battery power, solar irradiance, and PV panel selection while ensuring system reliability. The proposed system was optimized using improved Differential Evolution (DE) and its effectiveness was tested by comparing the results with Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). The Improved DE algorithm provides the highest average cost savings compared to other algorithms, which is $500 per year. It is recommended that this method is very useful in the optimization of off-grid PV systems, considering other uncertainties that affect PV system performance.