This article presents the real time implementation of modified neural network algorithm-based control scheme for three-phase grid connected solar photovoltaic (PV) systems. The modified neural network-based controller estimates the reference current, which is further required to produce pulses for the switching of Insulated Gate Bipolar Transistor (IGBTs) of Voltage Source Converter (VSC). The adaptive conjunction constant of the implemented control scheme is dimension less that makes the conversion step size of the implemented control, a data depended and time varying with a very high dynamic response.The purposed control is highly suitable for feeding intermittent duty loads. This is achieved with the help of very fast reactive power compensation capability of the implemented control during sudden and frequent variations in load current. The impacts of sudden and frequent changes in load current on the system due to intermittent duty loads are mitigated at adaption stage of the control. The VSC is supplying reactive power and harmonic currents required to the nonlinear loads. It also converts DC power of solar PV into AC power of desired voltage and frequency in synchronization with the grid. It supplies power demanded by the load and surplus power is pumped into the grid.artificial neural network, grid connected, neural-network, power generation, power quality, solar energy, solar power, VSC * , Reference grid currents; i vsca , i vscb , i vscc , Three-phase VSC currents; L 1 , Interfacing inductor; P PV , Solar PV power; S 1 ,S 2 , S 3 ,S 4 ,S 5 ,S 6 , Gating signal; V dc , Sensed DC link voltage; V dc * , Reference DC link voltage; V PV , Solar PV voltage; v sa , v sb , v sc , Sensed voltages at PCC.