On-line quantitative analysis of slag, which could greatly improve product quality and reduce energy consumption, is a highly demanded technique in metallurgic industry.Laser induced breakdown spectroscopy combined partial least squares regression model was proposed to determine the content of CaO, MgO, Al2O3 and Fe in slag.Background correction and spectral normalization, which used plasma intensity as reference signal, were applied to improve spectral signal stability.5 slag samples were analyzed by using the partial least squares regression model established with 25 slag elements-known samples.The mean prediction relative error of CaO, MgO, Al2O3 and Fe was 4.7%、11.5%、17.9% and 12.5%, respectively.The experimental results indicate that laser-induced breakdown spectroscopy combined PLS is a potential tool for on-line quantitative analysis of slag.