This study presents a novel approach to estimate the maximum power point (MPP) of a photovoltaic (PV) array on an hourly basis using meteorological data. The estimation of MPP of a PV array based on its characteristic equation is done by using Levenberg-Marquardt (LM) method. The monthly averaged daily and hourly global solar irradiation during sunshine hours of Guwahati city is estimated by taking relative humidity and ambient temperature data using empirical relations. The actual and estimated maximum powers of a PV array of monthly averaged daily and hourly basis is calculated using MATLAB. It is demonstrated that each of the estimated maximum power of a PV array on an hourly basis is nearly equal to each of the actual maximum power. The MPP of a PV module estimated by the LM method is validated with the actual and experimental results. A comparative study of MPP of a PV module estimated from the proposed method with perturb-and-observe based on proportional-integral, genetic algorithm and Gauss-Seidel is performed. The performance of the proposed algorithm for estimation of MPP of a PV array ensures the accuracy and faster convergence. © The Institution of Engineering and Technology 2016.