This paper introduces a new two-stage maximum power point tracking (MPPT) technique for multiple photovoltaic arrays operating under different levels of irradiance and temperature. The variations in such conditions could cause several local maxima on the overall power-current (P-I) curve of the arrays. This technique aims to locate the global maximum power point (MPP) on the P-I curve of the interconnected arrays thus bypassing any local maximum that might trap available single-stage MPPT algorithms. The first stage of the proposed technique finds a point that bypasses local maxima and moves the operating point of the PV arrays near the global MPP. The second stage is a normal MPPT technique that finds the exact global maximum. A system of two series PV arrays, a battery load, and the proposed technique were simulated in Simulink. The operation of the proposed two-stage technique is verified with the second stage being either the ldquoperturb and observerdquo algorithm (P&O) or the ripple correlation control (RCC). The efficiency of the technique is around 95% and its convergence time is 11 ms under extreme changes in the operating conditions. Compared to single stage techniques and other two stage techniques, this technique is shown to be very competitive, accurate, and fast.