Endpoint prediction of copper converter blowing is significant to entire copper pyrometallurgical process. It provides effective information for controlling production and the prevention of over-blowing and under-blowing is essential to ensure product quality. To address the multivariate coupling, nonlinearity and uncertainty characteristics caused by the intermittent cyclic operation of copper converter blowing, we propose the FA-ABC-RELM model to achieve accurate prediction of copper converter blowing endpoints. Factor analysis (FA) is used to reduce the raw data dimension and extract key information. Then, regularized extreme learning machine (RELM) model optimized by artificial bee colony (ABC) algorithm is used to predict blowing endpoint. It improves the model stability, prediction accuracy and generalization ability through the introduction of regular terms and parameter optimization. Simulation experiments are conducted with the production data of copper plant, and the proposed method is comprehensively analyzed in combination with the back propagation (BP) neural network model. The experimental results show that the FA-ABC-RELM model has higher fitting degree and smaller prediction error. Therefore, the developed method can be implemented as a promising endpoint prediction method for copper converter blowing, and it provides a guiding significance in quality control and correction of copper smelting process.