The introduction of generative artificial intelligence is of unique help to the field of government environmental governance. Generative artificial intelligence can quickly organize environmental data. As a digital customer service officer, it can realize humanized expression, assist staff to make environmental decisions and strengthen environmental supervision. The deep application of generative artificial intelligence embedded in government environmental governance has driven a logical change in governance subject, time, method and content. However, there are different risks in the early preparation stage, the actual operation stage and the subsequent treatment stage when the generative artificial intelligence is embedded in government environmental governance. Therefore, it is necessary to introduce the theory of agile governance, integrate the comprehensiveness, adaptability and flexibility of agile governance into the design of regulatory path, and construct an adaptive governance mechanism that conforms to the basic principles of the comprehensiveness of agile governance, the combination of prevention in advance and response afterwards, and the flexible process supervision mechanism.