In large-scale photovoltaic generators, the arrangement of modules with different electrical characteristics could involve a considerable mismatch between the single components resulting in a power loss. This means the actual power is less than the sum of the maximum output powers of the individual PV modules, operating at same irradiance-temperature conditions. Tn reduce the mismatch losses and to calculate it under operating conditions, a statistical approach based on Monte Carlo simulation techniques, has been developed and validated. The simulation model shows that it is possible to meet the required mismatch level, with a random arrangement, starting from a modules population characterized in terms of short circuit current, I-sc and open circuit voltage V-oc, by a probability density function with a imposed variance. The method has been successfully applied for a 100 kWp standard unit photovoltaic generator, the computational results have shown good agreement with the experimental data.