The traditional requirement modeling method describes requirements through a formal modeling language, which is difficult to adapt to the dynamically evolving requirements in the Cloud ERP domain due to its lack of flexibility and supporting requirement solutions. Aiming at the above problems, a Cloud ERP modeling and evolution method based on the E-CARGO(Environment-Class, Agent, Role, Group, and Object)model was proposed, relying on the feature that E-CARGO can flexibly model the evolving requirements. Firstly, the requirements were deconstructed into needs and Cloud ERP services and described as a role-based collaboration problem, which was then modeled by the E-CARGO model; secondly, in order to conduct more domain-specific assessment of talent competencies in the Cloud ERP domain and to build high-quality Cloud ERP development teams, a fuzzy complementary judgment matrix was established using the Fuzzy Analytic Hierarchy Process(FAHP)to measure talents' comprehensive ability value; finally, the group role assignment method using E-CARGO model combined with CPLEX optimization package was used to obtain the requirements maximization satisfaction scheme in all evolutions. The experimental results show that the proposed method is not only able to model the evolving requirements in Cloud ERP flexibly, but also can find the requirement satisfaction solution in a very short time, and then respond to the evolving requirements quickly and provide Cloud ERP services on demand.