Abstract: Due to the mass concentration of component was hard to measure online in hydrometallurgy extraction process，so a data modeling method was proposed by combining independent component analysis with least squares support vector machine(ICALSSVM)， which was used to establish the forecasting model of copper extraction.First， the independent component analysis was used to pretreat the field data. Then， the least squares support vector machine was used to established distribution ratio model which was used to identify the unknown parameter and enhance the accuracy of model. The simulation results showed that the models were practical and efficient， which could provide a theoretical basis for the control.

Key words: extraction, hydrometallurgy, hybrid model, ICA, LSSVM