This manuscript presents current sensorless algorithms for maximum power point tracking (MPPT) in partially shaded photovoltaic (PV) systems. The necessity of a current sensor is eliminated with the use of mathematical modeling of the power electronics converter. This approach significantly reduces the implementation cost and the inherent disadvantages in the current sensor circuitry. MPPT techniques based on soft computing are employed, in addition to Perturb and Observe (P&O), due to their ability to explore a larger search space. This feature is advantageous because it minimizes convergence risk to a local maximum, a limitation of traditional techniques. Simulation and experimental results are presented and each algorithm is evaluated through different metrics, such as search time for the global maximum power point (GMPP) and efficiency. The tests consider dynamic irradiance profiles, producing a tracking factor (TF) above 99% and a remarkable fast convergence time.