Cohesion is the key cure factor in group psychotherapy research area, which is the counterpart of the therapy alliance in individual psychotherapy. At the beginning of group cohesion research, cohesion is defined as attraction-to-group and interpersonal attraction, which is operationalized as interpersonal trust, attraction and involvement. With more and more understanding on cohesion, researchers began to combine different aspects to explore cohesion in multidimensional perspective. There were mainly three angles of view. One was about the structure of cohesion, based on the analysis to member-member, member-group, and member-leader relationships. Another was about the content of cohesion, such as Stokes’ viewpoint, he thought that cohesion in therapy group include the group attraction to individual member, the instrumental value of the group, and risk-taking behaviors that occur in the group; Braaten proposes a 5-factor-model of group cohesion: attraction and bonding, support and caring, listening and empathy, self-disclosure and feedback, and process performance and goal attainment. There was still a third view that exploring cohesion from both the structure and content, for example, Johnson found a three-dimensional-structure-model of cohesion, they are positive relationship, positive working relationship, negative relationship, respectively. Hornsey put forward that the social psychology theory can be integrated into group psychotherapy research, based on social identity theory and other theories, he proposed that "group identification", "in-group homogeneity", "task interdependence" and others can replace the conception of cohesion, but this construct still need further empirical support. Social identity theory provides a good perspective to explore cognitive level of group cohesion, the traditional view of group attraction lays the foundation for emotional level of cohesion, and goal congruence is helpful for exploring motivational level, all of above provide a new possibility for developing cohesion conception through the integration of these theories. There are many kinds of cohesion measurement tools developed by foreign researchers, but theories no consistency between them which result in the cohesion studies lacking comparability which leads to disadvantage influence on further research. At present, the group climate questionnaire-short form (GCQ-S), therapeutic factors inventory-cohesiveness scale (TFI), and cohesion to the therapist scale (CTS) are the tools recommended by American group psychotherapy association (AGPA) to evaluate the group process among those standardized measures to evaluate the clinical effect However, there is no consensus on the definition of cohesion, and systematic theory and model is also undeveloped. Domestic group psychotherapy researches still remain in effect research stage. It is worth domestic researchers concerning about the direction that exploring group psychotherapy process, developing assessment tools, building systematic theory and model.