Abstract: An effective scientific research evaluation system is the basis for the development of scientific research innovation activities in universities, and it can promote the improvement of scientific research innovation capabilities. Besides, the construction of "Double First-Class" in China puts forward new requirements for the existing scientific research evaluation system of universities. As a country with early reform of scientific research evaluation system in universities, the United Kingdom has rich and effective experience.Based on the above considerations, this paper summarizes the reform background, evolution process and system framework of scientific research evaluation in British universities firstly. After three stages of development, the British scientific research evaluation system has formed a scientific research excellence framework (REF). The REF is mainly composed of five parts, including evaluation unit, evaluation index, evaluation level, evaluation method and evaluation results.Then, the characteristics of REF are analyzed, including evaluation index, evaluation unit, evaluation method and research output. Results indicate that REF focuses on the innovation of scientific research achievements, the social value and social influence of scientific research achievements. On this basis, the differences between the scientific research evaluation system in Chinese universities and the British REF evaluation system are summarized; thus, the improvement suggestions for the scientific research evaluation system in China is further put forward from four aspects: discipline characteristics, third-party evaluation institutions, quality of scientific research achievements and interdisciplinary development.Finally, the reform idea of scientific research evaluation system in China is put forward, combining with the era background of "Double First-Class" construction, which are as follows: (1) Assessment team and assessment unit settings. The evaluation team is set up according to different disciplines and majors, and the first-level discipline is used as the evaluation unit for classification evaluation. (2) Evaluation level settings. Five levels can be set for the evaluation of scientific research achievements: international leading, international recognition, domestic leading, provincial leading and not recognized. (3) Evaluation cycle settings. The normal cycle of scientific research evaluation in colleges and universities is 5 years. (4) Evaluation basis. Due to the differences between disciplines, the assessment team of each assessment unit can formulate assessment standards for their own disciplines on the premise of the three basic indicators. (5) Universities have the right to choose independently. Universities can decide on their own the type and number of first-level disciplines to participate in the evaluation, and independently determine the scholars participating in the scientific research evaluation and their representative achievements. (6) Correlating evaluation results with funding allocations. For the use of the evaluation results, scientific research funds such as "Double First-Class" discipline construction funds, national key laboratory operation funds, key discipline construction funds, and basic scientific research business fees in central universities can be centrally managed, and allocated according to the evaluation results.In summary, the evolution process and system framework of the system framework and characteristics of scientific research evaluation in British universities are summarized, and its reference significance for China are analyzed in this paper. On this basis, the reform idea of scientific research evaluation system is proposed. This study can provide critical insights for the scientific research evaluation in "Double First-Class" construction.

Key words: scientific research evaluation in British universities, system characteristics, reference for scientific research evaluation, evaluation mechanism reform