Conventional translation systems are inevitably limited in terms of accuracy, fluency, and flexibility. This study is an effort to design and implement a diversified AI translation system based on syntax and semantics to improve the quality of translation. In this study, it uses the syntax and semantics methodologies to analyze and align the syntactic structure in target and source language, model and represent semantics and the relationships between the words in order to achieve translation with more accurate, natural and understandable meanings. At the same time, multiple models and algorithms of translation are proposed, and through trade-offs and choices, a variety of translation options and styles are achieved. The diversified AI translation system based on syntax and semantics improves the quality, performance, accuracy, and flexibility of translation in the traditional sense. The diversified AI translation system based on syntax and semantics achieves practice results that improve the quality of translation and the flexibility and efficiency of the translation system. In practical application scenarios of cross language information exchange, the diversified AI translation system based on syntax and semantics has made the existing technologies have a positive effect on cross language communication and information exchange. This is also an important aspect of the evolution of translation technology.