With the arrival of the low carbon era, enterprises, as the main body of the market economy, must take the road of low carbon operation whether to fulfill their social responsibilities or to seek enterprise development. However, Chinese enterprises started late to understand the low-carbon economy. So far, except for a few state-owned enterprises, some leading private enterprises and foreign-funded enterprises, they have made remarkable achievements in low-carbon emission reduction. Most enterprises in China have not realized the importance of low carbon emission reduction. In order to step out of the “high carbon” era and coordinate economic development and environmental protection, China must solve the problem of low-carbon transformation of small and medium-sized manufacturing enterprises. How to measure the degree of low carbon transformation, and how the government objectively evaluates the degree of low carbon economic development of enterprises, with a view to formulating corresponding standard incentives and punishment measures for them. It is urgent to establish a sound, reasonable and feasible low carbon operation management indicator system, and comprehensively evaluate the information related to low carbon operation of SMEs through reasonable selection of indicators. According to the connotation of low-carbon enterprises, the economic benefits of enterprises are closely combined with environmental benefits. A set of low carbon operation performance evaluation index system including economy, technology and environment has been constructed. The AHP index weight is calculated by establishing the AHP model and the Data Envelopment Analysis (DEA) index weight is calculated by establishing the DEA model. On this basis, the grey relational evaluation model based on AHP-DEA is established. Then the paper evaluates the performance of representative enterprises’ low carbon operation. Propose corresponding improvement suggestions for the problems reflected in the low carbon operation performance ranking of enterprises and the scoring of various indicators. The research results show that the algorithm has high efficiency and the accuracy of model evaluation is 95.51%. To a certain extent, this study makes up for the research results of the impact of digital transformation on corporate strategic performance, and also provides ideas for the research of corporate financial performance evaluation.