The employment of vocational college graduates is a matter of social stability. In this paper, we first introduce modularized teaching in vocational college students’ employment guidance courses to unify the teaching of vocational college students’ career development and employment guidance courses. Secondly, in the preference analysis module, the WOA algorithm is used to optimize the selection of LSSVM parameters, and the WOA-LSSVM model adapted to the regression prediction of vocational college students’ employment preferences is proposed. Then, in the module of employment optimal matching, the Lagrangian conditional function extreme value is found in the discrete job data through the residual optimization algorithm, and the results of the optimal matching of vocational college students’ employment supply and demand are obtained. Finally, the overall situation and group differences in the intensity of employment pressure felt by vocational college students are analyzed. The results show that the maximum value of students’ employment pressure in the modular teaching of employment guidance is 128.28, the mean value is 0.02283, and the Std.Dev is 0.90383. The multivariate kurtosis coefficient decreases from 6.385 to 0.874, which is the critical value (c.r.). Decreases to 1.983<1.95, and the data are normally distributed. This study provides some references to enhance the career planning function of vocational college students’ career guidance courses.