To explore the characteristics of the brain function network in treatment-naive young and middle-aged patients with depression. Methods: Twenty-four patients with major depression and 25 healthy volunteers were enrolled in this study. The severity of depression was assessed using the Hamilton depression scale (HAMD). All participants underwent DTI scanning, their brain networks were constructed using the graph theory method, the relevant network indexes were calculated, and the modules were estimated by Gretna's method. Finally, the brain network attributes and modular characteristics of the two groups were compared. Results: Compared with the healthy control group, patients with major depression had higher homozygosity (t=2.391, P=0.024); higher median values of left supplementary motor area and right superior temporal gyrus (t=2.564, P=0.016); and lower node efficiency of the left superior frontal gyrus, left middle temporal gyrus, left middle occipital gyrus, and left paracentral lobule. Further, the shortest path of nodes in the right amygdala, left insula, and left superior marginal gyrus was decreased in patients with major depression (t= 2.038, P=0.049; t=2.174, P=0.046; t=2.262, P=0.032; t=2.043, P=0.048), and edge connectivity analysis showed that the edge connectivity of the right caudate nucleus and the right lenticular putamen in patients with major depression was lower than that in healthy controls (P=0.000; P=0.000). Finally, we identified negative correlations between the median value of right superior temporal gyrus and Hamilton Depression Scale scores (r =－0.635, P=0.012). Conclusion: The brain network of patients with major depression is abnormal; the functions of some brain regions related to major depression are altered and show recombination, for which local brain regions may compensate.