Building a high-quality English corpus is a prerequisite for implementing natural language processing technology, and corpus alignment is an important part of natural language processing technology. However, currently, corpus alignment suffers from poor domain adaptability and alignment accuracy due to significant language differences and imbalanced data volume. In response to the issue of corpus alignment, this paper applies Particle Swarm Optimization (PSO) algorithm to the construction of English corpus, and combines corpus annotation for alignment feature analysis. The experimental results show that the English corpus constructed in this article can effectively solve the problem of word vector mapping in natural language processing, and the PSO method has significant advantages in alignment accuracy, resource consumption, and operational efficiency. The highest accuracy can reach 97.5%, indicating that the application of PSO effectively improves the performance and accuracy of alignment systems in natural language processing.