Objective To explore the predictive relationship between pressure perception and loneliness of college students by longitudinal study. Methods From October 2018 to June 2019， senior college students（ juniors and seniors） from three universities in Heilongjiang Province were selected as the research objects by convenience sampling method. Three questionnaires were conducted with an interval of 4 months. The first test was conducted in October 2018（ T1）， and 912 valid subjects were obtained. The second test time（ T2） was February 2019， and 873 valid subjects were obtained. The third test time（ T3） was June 2019， and 806 valid subjects were obtained. The sample adopted in the study was 742 students who participated in all the tests. A oneyear follow-up study was conducted on the 742 college students with Chinese Perceived Stress Scale（ including two dimensions of tension and out of control） and UCLA loneliness scale. The data were analyzed by repeated measurement variance， Pearson correlation analysis and cross-lagged regression analysis. Results （1） The variance analysis results showed that there were significant differences in pressure perception and loneliness before and after measurement（ F=4.85， P＜ 0.05； F=29.65，P＜ 0.05）， but no significant differences in gender （F=0.06，P＞0.05； F=0.41，P＞0.05）（. 2） Correlation analysis of the three times of survey showed that there was a significant positive correlation between stress perception and loneliness（ P＜0.01）；（ 3）Cross-lagged regression analysis indicated that T1 pressure perception significantly predicted T2 pressure perception（ β=0.45，P＜0.01） and T2 loneliness（ β=0.17， P＜0.01）. T1 loneliness had a significant predictive effect on T2 pressure perception （β=0.22， P＜0.01） and T2 loneliness（ β=0.59， P＜0.01）. T2 pressure perception had a significant predictive effect on T3 pressure perception（ β=0.47，P＜0.01） and T3 loneliness（ β=0.21， P＜0.01）. T2 loneliness had a significant predictive effect on T3 pressure perception（ β=0.20， P＜0.01） and T3 loneliness（ β=0.51， P＜0.01）. T1 tension had a significant predictive effect on T2 tension（ β=0.35， P＜0.01）， but not on T2 loneliness（ β=0.04， P＞0.05）. T1 loneliness had a significant predictive effect on T2 tension（ β=0.18，P＜0.01） and T2 loneliness （β=0.67， P ＜ 0.01）. T2 tension had a significant predictive effect on T3 tension（ β=0.35，P ＜ 0.01） and T3 loneliness（ β=0.10，P＜0.01）. T2 loneliness had a significant predictive effect on T3 tension（ β=0.21，P＜0.01） and T3 loneliness（ β=0.59， P＜0.01）. T1 loss of control had a significant predictive effect on T2 loss of control （β=0.37， P＜0.01） and T2 loneliness（ β=0.14， P＜0.01）. T1 loneliness had a significant predictive effect on T2 loss of control（ β=0.22， P＜0.01） and T2 loneliness（ β=0.63， P＜0.01）. T2 loss of control had significant predictive effect on T3 loss of control（ β=0.33， P＜0.01） and T3 loneliness（ β=0.14， P＜0.01）. T2 loneliness had a significant predictive effect on T3 loss of control（ β=0.15， P＜0.01） and T3 loneliness（ β=0.57， P＜0.01）. Conclusions There is a causal relationship between pressure perception and loneliness among college students.