Background: In this paper, a new hybrid global maximum power point tracking (GMPPT) technique is proposed for faster and accurate tracking of global maximum power point (GMPP) without premature convergence. It is a combination of modified particle swarm optimization (PSO) and perturb and observe (P&O) methods. The proposed GMPPT technique, adaptive butterfly PSO (ABF-PSO) uses butterfly swarm intelligence for modifying the conventional PSO algorithm with parameter tuning to avoid premature convergence. Aims: Hybrid global maximum power point tracking (GMPPT) technique is proposed for faster and accurate tracking of global maximum power point (GMPP) without premature convergence. Further, a new reinitialization of particles for any irradiance change is proposed to get faster tracking. Materials & Methods: In the proposed hybrid GMPPT technique, first GP region is easily identified with adaptive sensitivity parameter of the ABF-PSO algorithm and in the region identified, GMPP tracking is continued with P&O algorithm with variable length perturbations to avoid the unnecessary exploration of search space even after reaching global peak (GP) region. Results: The combined effect of adaptive parameters, global region identification with adaptive sensitivity, proposed reinitialization method, and steadystate tracking with variable step P&O results in fast and accurate tracking of GMPP with low power oscillations during GP region identification stage and steady-state.