Abstract: Due to the high content of corrosive substances in the oil-gas-water mixed pipeline, the corrosion rate is relatively fast. This research proposes a corrosion rate prediction model based on KPCA-ICS-ELM model. First, the KPCA（nuclear principal component analysis） algorithm, ICS（improved cuckoo search） algorithm, and ELM（extreme learning machine） algorithm are introduced respectively. The KPCA-ICS-ELM model combination method and model evaluation method was proposed. On this basis, the corrosion rate data of the oil-gas-water mixed transportation pipeline was obtained through experiments, and then the index system was constructed. The feasibility of the proposed model is compared with other models to prove the advanced nature of this model wasproposed. The research shows that by using the KPCA model to analyze the influencing factors that affect the uniform corrosion of the inner wall of the mixed transportation pipeline, it is found that the H₂S content, CO₂ content, temperature, flow rate, and pH value in the medium have a relatively large impact on the uniform corrosion rate. The minimum error of the model proposed by the research institute for uniform corrosion rate prediction is 1.24%, the root mean square error is 0.9339%, and the Hill variability coefficient is 0.5273%. Compared with other models, the accuracy of the model proposed in this research is relatively high, which proves that the use of the KPCA-ICS-ELM model to predict the uniform corrosion rate of oil-gas-water mixed pipelines is relatively feasible and advanced.