The study proposes a dynamic resource allocation model suitable for English language teaching, which combines learner characteristics, learning progress and resource availability to achieve real-time optimal allocation of resources through mathematical optimization algorithms. A multi-objective optimization model is constructed based on the key factors in resource allocation for English teaching. Facing the optimization objectives of maximizing learning efficiency and minimizing resource idleness, NSGA-II algorithm is used to construct a non-dominated solution to achieve global sorting, and combined with congestion calculation to complete global quality population screening. At the same time, the branch delimitation algorithm is utilized for local search of optimal solutions, and merged with the population of NSGA-II to generate the new generation of optimal populations. The optimization probability of the combined algorithm in this paper is 0.85, and the average convergence error is only 0.01081, which has excellent optimization performance. The resource allocation delay of this algorithm is around 0.1ms, and the allocation efficiency is more than 95%, and the comprehensive effectiveness is better than the comparison algorithm. The dynamic allocation model of resources in this paper improves the balance of resource allocation of English teaching and auxiliary room area, the number of teaching materials, the number of full-time teachers and teaching equipment. At the same time, it prompted the average English score of the experimental class to exceed 80, which was significantly higher than that of the control class.