Abstract Exploring the construction, of course, Civics of graduate students in foreign languages is to promote better the implementation of course, Civics in foreign languages. In this paper, in the context of the new liberal arts of big data, firstly, the method related to the sparrow search algorithm is described, and the parameters related to the sparrow search algorithm are optimized by using the good point set initialization and firefly update strategy. Then the basic principle of the extreme learning machine is given, and it is shown that the ELM model’s output weights are obtained using the inverse matrix solution. The ISSA-ELM algorithm model is elicited using the improved sparrow search algorithm with the extreme learning machine, and the ARE and RMSE performance validation is performed for this model using the UCI dataset. Finally, the ISSA-ELM algorithm was used to analyze and mine the indexes of the construction of the example of the postgraduate course Civics in foreign languages, including both teaching implementation and teaching content, using GZ University as an example. From the teaching implementation, the percentages of A, B, C and D evaluations are 16.12%, 18.44%, 18.12% and 47.32%, respectively. Regarding teaching content, the percentages of A, B, C, and D evaluations were 49.65%, 22.95%, 16.5%, and 10.89%, respectively. This indicates that the construction of the course Civics for graduate students in foreign language disciplines in the context of the new liberal arts is not optimistic, and effective reforms in teaching implementation and content are needed afterward.