In this paper, an efficient active power filter (APF) scheme is developed to estimate and compensate for harmonic distortion in an electrical power network. The developed APF control scheme is based on a double proportional feedback controller and a single-phase voltage-source half-wave bridge inverter. The proposed filter uses a Multi-layer Artificial Neural Network (MLANN) with a shift method for estimating system harmonic currents and voltages at a dedicated point. The proposed scheme is tested on a 13 bus industrial distribution system. The obtained results ensure the effectiveness of the proposed filter. The system is simulated in MATLAB-Simulink and simulation results prove that the polluting harmonics have been greatly reduced.