In this work, we present a comprehensive combined modelling approach to study the annealing of lattice defects in dilute and concentrated metallic alloys. The developed approach consists in the combination of molecular dynamics, atomistic kinetic Monte Carlo (AKMC) and mean field rate theory methods, linked at appropriate time and space scales. For the first time, the AKMC tool has been designed to model the evolution of point defects (both vacancies and self-interstitial atoms) in random concentrated alloys, taking into account the influence of lattice distortion on the local migration energy barrier due to the mutual interaction of point defects and solutes. Good accuracy and outstanding speed of calculations has been achieved by introducing the artificial neural network regression as an engine of the AKMC applied to calculate migration barriers for mobile defects. The developed method was applied to study correlated recombination in bcc Fe and random Fe–Cr alloys, aiming at the reproduction of a set of experimental studies after electron irradiation. The obtained results agree well with the available experimental data, implying that the developed modelling procedure correctly captures the undergoing physical process.