Environmental judicial adjudication is an essential component of the eco-environment governance system. The artificial intelligence large language model (AI-LLM), developed based on generative artificial intelligence, has offered significant opportunities for the environmental judicial adjudication to develop toward a higher level of smart adjudication. This study aims to promote the integration of AI-LLM technology with environmental judicial adjudication and promote the intelligent development of the environmental judicial adjudication. It explores the role and practical applications of AI-LLM in environmental judicial adjudication and summarizes prominent problems such as poor data quality, bias caused by algorithmic opacity, and limited capabilities for deep application. Using eco-environmental protection cases as an example, this study establishes a smart environmental judicial adjudication system based on AI-LLM and elaborates on the architecture design of the system and the technical elements involved. Furthermore, it proposes the following suggestions to promote the intelligent development of environment judicial adjudication: (1) emphasizing top-level design and establishing a high-end think tank for environmental justice; (2) building an environmental justice data center to improve the judicial data standards system; (3) establishing an algorithmic governance mechanism to promote the fairness and justice in environmental judicial adjudication; and (4) improving the multiple accountability mechanism for environmental justice to strengthen the judicial supervision and management system.