 This paper proposes a novel approach combining particle swarm optimization, PSO, and artificial neural networks, and to accurately identify individual appliances in smart homes. The PSO algorithm was used to optimize the parameters of the ANN model, while the ANN was trained to distinguish between different appliances based on their power consumption. The proposed approach was tested on three datasets and compared against existing methods. The results showed that the proposed approach achieved higher accuracy than the existing methods, with an average NIMSE of 1.719%. Additionally, the proposed approach was able to accurately predict customer behavior regarding energy usage during daytime hours. This article was authored by R. Ramadan, Qi Huang, Alusola Bamiasile, and others.