The Neuro-fuzzy models have been used and applicable in different fields of electrical engineering, industrial ergonomics, science and given a good result. This study presents a control method for grid-connected photovoltaic (PV) systems based on a a Adaptive Neuro-Fuzzy Inference System (ANFIS) controller. Multi-Level Inverters (MLIs) have become very popular in industrial and grid-connected applications in recent years due to their multiple advantages. Integrate PV power into the power grid with a 9-level Neutral Point Converter (NPC) to minimize harmonic distortion and maximize efficiency. The output voltage of a solar array varies greatly with solar irradiance, but an inverter output voltage must be maintained to connect to the grid. To achieve this, To do this, three ANFIS controllers were used to control the inverter output voltage (Vdc), direct current (Id) and quadratic current (Iq) around a reference value. A comparison is made with the PI regulator. According to research, ANFIS controller outperforms PI controller in performance

ANFIS controller - PI controller, inverter, MLI, PV array, THD;