A chloride process is especially prone to leakage incidents via vaporizer rupture and pipeline perforation, both of which are usually caused by corrosion. Rupture accidents may not occur frequently but nevertheless can results in explosions, poisoning and other severe consequences. The main factors affecting the corrosion process are the chloride ion concentration and the environmental temperature. In order to study the probability of rupture accidents, this paper first presents experimental research on the corrosion performance of liquid chlorine vaporizers and pipelines made with the steel Q345 and 20G, respectively. Different chloride ion concentrations (1-20 g/L) and environmental temperatures (30-90 degrees C) were considered in the corrosion experiment. The experiment lasted for 42 days, and the corrosion rate was calculated using weight loss technique. The experimental results show that 20G steel suffers more serious corrosion than Q345 steel and the degree of corrosion is determined by different factors working together. Among these experimental data, the corrosivity data related to 20G steel at the corrosion environment of low temperature with low concentration (30 degrees C, 1 g/L), as well as at high temperature with high concentration (80 degrees C, 20 g/L) were selected to establish the corrosion weight loss prediction model, based on which the accumulated wall thickness reduction was further calculated. Finally, a probability prediction model for the corrosive state was established based on Markov chains. Using the dynamic prediction method proposed in this work, it is capable to determine the corrosion situation of the chloride process through predicting the rupture probability before the coming into service of process equipment. Also, these prediction results can provide some guidance for the maintenance and help to reduce the risk of rupture accidents caused by corrosion in a chloride process.