With the increasing importance of English proficiency in global communication and professional settings, the need for effective English language training systems has become paramount. In the context of English listening skills, traditional training methods often lack adaptability and personalized learning experiences, resulting in suboptimal outcomes for learners. In response, this paper presents a novel approach that leverages machine learning techniques to assess the effectiveness of adaptive English listening training. This study introduces an IPSO-BP network as a means to assess the effectiveness of adaptive English listening training. In addressing the limitations of the particle swarm optimization algorithm, enhancements have been made from two aspects. Additionally, the proposed IPSO algorithm is employed to optimize the initial parameters of the BP network and establish an IPSO-BP network to overcome the issues inherent in the BP network. The effectiveness of adaptive English listening training is then evaluated through the utilization of the IPSO-BP network. A comprehensive and systematic experimental analysis was performed, confirming the accuracy and dependability of the approach. The experimental data affirms the correctness and reliability of this method in improving adaptive English listening training.