文章梳理了国内外潜在科研合作机会识别相关成果，归纳总结现有识别方法及存在的问题，为学科领域进行前瞻性合作推荐提供参考借鉴。首先对潜在科研合作机会识别的必要性进行归纳总结，其次对相关概念及研究主体类型进行界定，再次在调研国内外相关研究基础上对潜在科研合作机会识别方法进行归纳整理，最后指出现有研究不足并对未来发展提出展望。研究发现：就研究主体类型而言，根据研究目的、研究层次的不同，将研究主体划分为微观、中观、宏观三个维度。就识别方法而言，外部属性信息是潜在科研合作机会识别方法中最直接、最通俗易懂的方法；链路预测是使用最多、应用最为成熟的方法；比较而言，网络学习和机器学习是潜在科研合作机会识别的新方向和新思路。在以上分析基础上，总结了不同方法的不足以及存在的普适性问题，并对未来研究重点进行展望。

This Paper sorting out the relevant achievements in the identification of potential cooperation opportunities, this paper summarizes the existing identification methods and problems, providing reference for forward-looking cooperation recommendations in the discipline field. Firstly, this paper summarizes the necessity of identifying potential cooperation opportunities. Secondly, it defines the relevant concepts and entity object types. Thirdly, it summarizes the identification methods of potential cooperation opportunities. Finally, it points out the existing research deficiencies and puts forward prospects for future development. As for the types of entity objects, according to the different research purposes and research levels, entity objects are divided into three dimensions: microscopic, mesoscopic and macroscopic. As for the identification methods, external attribute information is the most direct and easy method in the identification of potential cooperation opportunities; link prediction is the most widely used and most mature method; in comparison, network representation learning and machine learning are new directions and new ideas for identifying potential cooperation opportunities. Based on the above analysis, the deficiencies and universal problems of different types of potential cooperation opportunities identification methods are summarized, and the future research priorities are prospected.