Thermoelectric generator (TEG) systems are gaining much attraction due to utility in heat recovery, surface cooling, concentrated solar thermal, and sensor applications. In series–parallel configurations, TEG modules due to Heterogeneous temperature difference (HeTD) show the non-linear behavior. Due to non-uniform temperature distribution (NUTD), multiple maximum power points (MPP) appears on the P-V curve. It is crucial to drive the system at true global maximum power point (GMPP) among multiple MPP’s. Existing classical techniques exhibit slow tracking, low efficiency, and undesired fluctuation in output voltage transients. To address these shortcomings our control technique based on improved Moth Flame Optimization (IMFO) is employed for the maximum power point tracking (MPPT) control under dynamic operating conditions. Comparison of the proposed technique is made with other well-known meta-heuristic techniques including particle swarm optimization (PSO), cuckoo search (CS), Artificial Bee Colony (ABC), and recently developed Dragon Fly Optimization (DFO). The comprehensive case studies with statistical and quantitative analysis are performed to confirm the superior performance of IMFO for NUTD condition, fast varying temperature condition, and stochastic operations. To experimentally validate the performance of the IMFO algorithm a low-cost TEG emulator setup is designed. The IMFO based control is implemented on a low-cost microcontroller achieving effective real-time control application in hardware. The proposed IMFO algorithm attains up to 6 W more power and takes 59% less time to track and settle at GMPP with minimum fluctuation. Results also validate that IMFO extracts 5.2% more electrical energy in comparison to competing techniques. In light of comprehensive analysis, it is safe to conclude that the proposed IMFO performs excellently for TEG MPPT control.