Abstract The paper focuses on the cost-effective design of a grid-connected rooftop photovoltaic (PV) system in Poland. The work aims to find the optimal system configuration, indicated by the highest Net Present Value (NPV), for the scenario of Net-Metering (NM) and Feed-In-Tariff (FIT) support mechanisms. PVsyst software was used to simulate the performance of various system sizes, tilts of the modules, as well as types and capacities of PV panels and inverters. Financial analysis was performed, by treating PV generation data in an economic evaluation model. At first, it was necessary to determine the building’s electricity demand, using the bottom-up method. Preliminary analysis, based on generic components, allowed to estimate the optimal size range of the system. Detailed simulations were later conducted in the specified ranges. Results of the simulation showed that the system capacity-optimized for NM was nearly twice bigger than for FIT, but the NM support system proves to be more beneficial, as it generates significantly larger NPVs. The sensitivity analysis showed that the three parameters most affecting the NPV of the systems studied were: electricity prices, the building's electricity demand, and capital expenditures.