In the process of cement production in the industrial field, the content of each component in the cement directly affects the quality of the cement. Therefore, it is of great significance to quickly and accurately monitor the content of each component in the cement. In this paper, the laser induced breakdown spectroscopy (LIBS) technology is used to detect the powder cement, and the powder cement are put in a two-dimensional moved platform without any pretreatment. The spectral data is processed by normalization and principal component analysis(PCA) firstly, which is used as the input of the model. In order to analyze the elements of Ca, Si, Al, Fe and Mg in cement, we build the models based on Partial least squares(PLS) and Support Vector Regression (SVR) as the comparison of methods. In addition, the comparison of measurement methods is between cement powder detection and cement tablet detection. The experimental results show that in this type of experiment, the SVR method is more advantageous than the PLS method because of the relationship between the element concentration and the strength of its characteristic line of the cement samples. The accuracy of the direct measurement of the cement powder is close to that of the tablet type, and it demonstrated the feasibility of on-line analysis of cement powder using LIBS technology under this type of experiment.