Abstract□□Objective: Change is the predominant goal of psychotherapy for all the theoretical orientations. Many researchers discuss the conditions of promoting the change from the angle of the clients, the consultants, and the counseling relationship variables. Ribeiro et al put forward an integrated framework and the conception of “Therapeutic Zone of Proximal Development” to explain the change process of the client. This study was aimed at exploring the characteristics of interactions in counseling of the good- and poor-outcome cases use the conception of “Therapeutic Zone of Proximal Development”, with the hope to guide the clinical practice. Methods: Based on the conception of “Therapeutic Zone of Proximal Development”, a Content Analysis was conducted to analyze the interactions in counseling of 5 good-outcome cases and 3 poor-outcome cases, employ the “Therapeutic Collaboration Coding System”. Procedures were as follows: (1) Preparation of the data. The data of the 8 cases of the talks were summarized, according to the requirements of the 1.5 round of talks as an analysis unit to form a text information assessment. (2) Simplification of the data: encoding and classification. After familiar with the “Therapeutic Collaboration Coding System”, each group assessed the type of therapist intervention, the type of client response, and the type of therapeutic exchanges for each assessment unit. Within the group to discuss the inconsistent encoding, and formation of a consistent encoding. (3) Data analysis and results presentation. Results: Following findings is revealed: (1) There are significant differences in the therapist interventions between the good- and poor-outcome cases (χ2=9.360, P=.002＜.05). Good-outcome cases use more supporting interventions and less challenging interventions than the poor-outcome cases do. Furthermore, there are also significant differences in the supporting interventions between the good- and poor-outcome cases (χ2=5.217, P=.022＜.05). Good-outcome cases support the dominant voice less than the poor ones, and support the no-dominant voice more than the poor ones. (2) On the client response, there is no significant difference between the good- and poor-outcome cases (χ2=.988, P=.320＞.05). The majority are validation responses, among which many cover-up invalidation may be coded as the validation and thus the ratio of validation responses is increased. This situation is more likely to occur in the cases of poor-outcome. (3) On the types of therapeutic exchanges, there are significant differences in the zone of “validation-safety” and “validation-tolerable risk” between the good- and poor-outcome cases, while there are no significant differences in the other zones. In the zone of “validation-safety”, good-outcome cases are less than the poor ones (χ2=7.639, P=.006＜.05), while in the zone of “validation-tolerable risk” good-outcome cases are more than the poor ones (χ2=13.528, P=.000＜.001). Conclusions: There are differences between the good- and poor-outcome cases. Clinical implications: counselors should consider to using more supporting interventions with appropriate combination of challenging interventions, increase support on the no-dominant voice. It could be helpful for counselors to carefully distinguish the cover components of validation response and adjust the interventions accordingly. In addition, it is suggested that counselors should “gently push the client forward along the ‘Therapeutic Zone of Proximal Development’”, and help clients eventually assimilate the no-dominant voices.