Assessment of the corrosion stage in reinforcing steel within reinforced concrete structures is crucial for determining the durability of such structures. However, research on the correlation between electrochemical indicators and the stages of corrosion experienced by reinforcing steel is limited. The primary impediment to this is the scarcity of study samples capable of simultaneously capturing data on half-cell potential (Ecorr), corrosion current (icorr), polarisation resistance (Rρ), and concrete resistivity (ρ) under specific structural corrosion conditions. It is essential to integrate the findings from multiple electrochemical indicators to determine the corrosion stage. Therefore, a Bayesian network is employed for the purpose of training to model the relationships among various electrochemical indicators and predict the missing data, thereby generating a comprehensive dataset. Subsequently, multilayer perceptron, support vector machine, K-nearest neighbour, naive Bayes, and decision tree models are trained and integrated using bagging resampling and soft voting techniques. Finally, a hybrid classifier system (HCS) is devised with Ecorr, icorr, Rρ, and ρ as inputs and used to determine the probability of reinforcing steel being in a particular corrosion stage. Based on the verification of the HCS model, it can be concluded that Ecorr has the greatest influence on determining the corrosion stage of reinforcing steel, followed by icorr and Rρ. When Ecorr, Rρ, and ρ are less than −242 mV, 209 kΩ cm2 and 2089 Ω m, respectively, and icorr is greater than 0.07 μA cm-2, the structure is considered to have entered the corrosion initiation stage and has begun to corrode. When Ecorr, Rρ, and ρ are less than −584 mV, 2.4 kΩ cm2 and 95 Ω m, respectively, and icorr is greater than 1.20 μA cm-2, the structure is considered to have entered the corrosion expansion stage. This study presents a quick and accurate method for the determination of the corrosion stage of reinforcing steel in RC structures.