Objective: This research aims to examine empirically whether there is arelationship between self-compassion and self-injury in first semester students,Faculty of Psychology, Unissula Semarang. Method: This research isquantitative using correlational methods. The research subjects were 88 firstsemester students consisting of 72 women and 16 men. The subject samplingtechnique used purposive sampling. This research data was collected using twoscales. The self-injury scale from Fiqria (2018) has an alpha reliability of 0.950,and the self-compassion scale from Susanti (2019) which has an alpha reliabilityof 0.920. Result: Research data analysis used non-parametric Spearmancorrelation, because the data distribution did not normal. The result obtained acorrelation coefficient of -0.437, and a significance level of 0.000 (p < 0.05)using Spearman's rho correlation. This correlation is stated to be quite strong(0.26 – 0.50 quite strong correlation category). Conclusion: It can be concludedthat the hypothesis is accepted, that there is a significant negative relationship between self-compassion and self-injury. The higher the student's self-compassion, the lower the student's self-injury, and vice versa.