The assessment of the shunt compensation allied power quality glitches bounds the design and modelling of a controller for the d-FACTS (Distribution Flexible Alternating current Transmission System) device. So, Kernel Hebbian least mean square (KHLMS) is proposed for controlling the distributed static compensator (DSTATCOM) in this study. The KHLMS is the improved version of an adaptive LMS (ALMS), which is designed by using a suitable pattern of learning mechanism. Both the controller algorithms are formulated on the basis of mathematical equations using MATLAB/Simulink. In accordance with the system variation adaptability and experimental application, each phase controller is configured as an independent neural network structure. The objective of control algorithms is used for the extraction of fundamental active and reactive components from load currents for the generation of the reference source currents. In comparison with the ALMS algorithm, the KHLMS demonstrates more effective in improving the voltage regulation at DC link capacitor, voltage balancing, source current harmonic reduction, and power factor correction under the disturbance influences caused by even and uneven loading. Nomenclature w pa , w pb , w pc weighting values of active component of load current w qa , w qb , w qc weighting values of reactive component of load current w a mean active weight w r mean reactive weight α learning factor γ step size i la , i lb , i lc load currents i sa , i sb , i sc supply currents u pa , u pb , u pc in phase unit voltage template u qa , u qb , u qc quadrature unit voltage template w cp , w cq output of proportional integral (PI) controller at both DC and AC sides w sp, w sq total active and reactive components of the reference source current i aa , i ab , i ac total active weight i ra , i rb , i rc total reactive weight v sa , v sb , v sc supply voltages i sa *, i sb *, i sc * reference supply currents