This manuscript proposes an Adaptive Linear Neural Network (ADALINE) based control for the unified power quality conditioner to protect the critical loads from voltage related power quality problems and mitigate the current related power quality problems produced by nonlinear loads. Due to parallel computing feature and simple algorithm, ADALINE is utilized widely in parameter estimation. In this manuscript, unified ADALINEs based current harmonic extractor for the shunt active power filter is utilized, and an ADALINE based control technique is utilized to produce reference voltages for load voltage compensation. Due to ADALINE control, the necessity of a phase locked loop and conventional filters are omitted, therefore, enabling the system to extract the fundamental components more accurately. Besides, the ability of self-regulating transition from islanded mode to grid-connected mode has ensured that uninterrupted power supply is provided to the critical loads. The system performance is validated through simulation results utilizing MATLAB-Simulink software considering several dynamic conditions, namely unbalanced grid voltage, self-regulating transition, and grid unavailability. © 2021 The Author(s)