The extraction of maximum power from solar photovoltaic (PV) using Maximum Power Point Tracking (MPPT) methods is a promising research area in the recent past. Many methods including conventional methods, such as Hill Climbing and Incremental Conductance, and methods based on neural network, Fuzzy logic and bio-inspired algorithms, were proposed for MPPT application. However, all these methods suffer from drawbacks such as slower convergence, reduced power output, predominant steady state oscillations, larger memory requirement, and complex structure. Hence, in this paper an attempt is made to enhance existing Particle Swarm Optimization technique by emphasizing proper initial value selection. The key features of this method include the ability to track the global peak power accurately under partial shading conditions with almost zero steady state oscillations, faster dynamic response, and easy implementation. Simulations are carried out for different shading patterns and the results obtained are compared with existing methods. Further, simulation results are validated via experimental values. © 2016 AIP Publishing LLC.