Objective To explore the effects of the two-different grouping methods on team-based learning（TBL） in Community Health Nursing. Methods The year 3 nursing undergraduates of two classes in a medical university from Fujian Province were selected by cluster sampling. The coin-tossing method was used to randomly divide the two classes into the experimental group（n=143） and the control group（n=135）. Two groups were both taught in TBL. Students in the experimental group were divided into small groups based on students’ grade-point average （GPA） following the“homogeneity between groups,heterogeneity within groups” principle. Students in the control group were divided into small groups based on their own willing. The differences in academic performance and teamwork and self-directed learning between the two groups before and after TBL application were compared. After the course,the evaluation of each group on the other group was collected. Results The academic performance of students in the experimental group（84.74±5.92） were significantly higher than that in the control group（82.10±6.39）（t=3.541,P<0.001）. The score of teamwork in the experimental group（24.09±8.77） was significantly lower than that in the control group（26.72±11.76） after TBL teaching（t=2.088,P=0.037）. The interpersonal communication dimen-sion of self-directed learning in experimental group（14.86±2.39） was significantly lower than that in the control group（15.70±2.07） after TBL teaching（t=3.093,P=0.002）. Conclusion “Homogeneity between groups,heterogeneity within groups” is more effective in enhancing students’ academic performance and improving their ability of self-discipline.