One of the major known challenges of improving durability of concrete structures is reducing the permeability of concrete to retard the transport of chloride ions. In this regard, the utilization of self-compacting concrete (SCC) incorporating supplementary cementitious materials such as fly ash and silica fume can aid to improve the resistance against chloride ion penetration. With the aim of evaluating the effectiveness of the SCCs, existing experimental methods, such as the rapid chloride permeability test (RCPT), demand time, financial resources, and skilled technicians. This study proposes three accurate and reliable predictive models through the combination of metaheuristic optimization techniques of particle swarm optimization (PSO), ant colony optimization (ACO) and biogeography-based optimization (BBO) with artificial neural network (ANN) in order to minimize the reliance on time-consuming and costly laboratory activities. To this end, 360 RCPT results on 60 SCC mixtures incorporating various percentages of fly ash and silica fume exposed to different temperatures during the RCPT were collected. The performance of the developed models was assessed by different statistical indicators. The results reveal that the proposed ANN-BBO model exhibits a higher accuracy than the single ANN, ANN-PSO and ANN-ACO models. It is also found that the input parameter of the sample temperature during testing is the most important variable in controlling the RCPT of SCC with a contribution of 25%. Comparing the performance of the best proposed model (ANN-BBO) with that of existing models in the literature reveals that the ANN-BBO model has a higher accuracy in predicting the chloride ion penetration resistance of SCCs.