Objective To analyze the morphological changes in the cerebral cortex of depressive disorder adolescents with non-suicidal self-injury（ NSSI） behaviors based on surface-based morphometry（ SBM） and its relationship with NSSI behaviors and clinical diagnostic value. Methods This study was prospective. From September 2022 to January 2024， 70 depressive disorder adolescents with NSSI behaviors（ nsMDD group） at Suzhou Guangji Hospital， China， were included for the assessment of NSSI function scale， addiction scale， and cranial magnetic resonance 3D-T1-weighted imaging scan. During the same period， 35 healthy volunteers were collected as a healthy control group（ HC group） and underwent MRI scanning. All subjects were analyzed using the Computational Anatomy Toolbox 12（ CAT 12） in the Statistical Parametric Mapping 12（ SPM 12） software package. Two independent samples t-test was used to compare the morphological parameters of the cortical surface in nsMDD group and HC group. Pearson correlation was used to analyze differential cortical surface morphometric parameters with NSSI function scores and addiction scores. The value of differential cortical surface morphometric parameters for diagnosing depressive disorder adolescents with NSSI behaviors was analyzed using the receiver operating characteristic（ ROC） curve. Results Compared with HC group， nsMDD group had a decrease in cortical thickness in the rostral part of the right anterior cingulate cortex ［（2.639±0.145） mm vs.（ 2.704±0.157）mm； t=2.090， P=0.039］， an increase in cortical fold index in the caudal part of the left anterior cingulate cortex［ （28.400±2.025） vs.（ 27.574±1.709）； t=-2.027， P=0.041］， a decrease in gyrus indices in the right supramarginal gyrus and the right transverse temporal gyrus［ （3.063±0.133） vs. （3.124±0.145），（ 4.706±0.270） vs.（ 4.862±0.167）； t=2.135， P=0.035； t=3.125， P=0.002］， and a decrease in the fractal dimensions of the left paracentral lobule［ （2.540±0.130） vs（. 2.606±0.081）； t=3.210， P=0.002］， and the differences were statistically significant. Correlation analysis showed that the cortical thickness of the rostral part of the right anterior cingulate cortex was negatively correlated with the NSSI function score， addiction score， and HAMD-24 score（ r=-0.570，-0.630，-0.401； all P ＜ 0.001）， and the gyrus index of the right transverse temporal gyrus was positively correlated with the NSSI function score（ r=0.248，P＜0.05）， and the differences were statistically significant. Multiple linear regression analysis showed that the NSSI function score （t=-3.267， P=0.002） and addiction score（ t=-4.628， P＜ 0.001） negatively predicted the cortical thickness of the rostral part of the right anterior cingulate cortex， and the difference was statistically significant. The ROC curve analysis showed that the area under the ROC curve predicted by the joint model of five metrics， including cortical thickness of the rostral part of the right anterior cingulate cortex， cortical fold index in the caudal part of the left anterior cingulate cortex， gyrus indices in the right supramarginal gyrus and the right transverse temporal gyrus， and fractal dimension of the left paracentral lobule was 0.773. Conclusions The cortical thickness of the rostral part of the right anterior cingulate cortex is closely related to the occurrence and severity of NSSI behaviors in depressive disorder adolescents， or may provide an important entry point for early clinical diagnosis of the disease.