Objective To explore the value of heart rate variability indices in distinguishing unipolar depression from bipolar depression. Methods This study adopted a cross-sectional survey. From January 2020 to June 2023， 126 patients with unipolar depression who were currently experiencing depressive episodes at the Shanghai Mental Health Center were selected as the unipolar depression group， and 247 patients with bipolar depression were selected as the bipolar depression group. Clinical Data Questionnaire and heart rate variability indices were used to compare the differences in clinical data and heart rate variability between patients with unipolar depression and bipolar depression. Logistic regression was used to evaluate the discriminative ability of heart rate variability indices in distinguishing unipolar depression from bipolar depression， while controlling for potential confounding variables such as age， gender， and body mass index， and to construct new individual models to predict probability variables. Results The body mass index of patients with bipolar depression was higher than that of patients with unipolar depression， and the difference was statistically significant （t= -2.798， P ＜ 0.05）. There was no statistically significant difference in the dicrotic elasticity index and low frequency/ high frequency between the two groups of patients （P ＞ 0.05）. The heart rate variability indices （SDNN， pNN50， very low frequency， low frequency， high frequency， total power） of patients with bipolar depression were lower than those of patients with unipolar depression， and the cardiovascular characteristic indicators （heart rate， ejection elasticity index， and dicrotic dilation index） were higher than those of patients with unipolar depression， with statistical differences （all P ＜ 0.05）. After controlling age， gender， and body mass index， there was no statistically significant difference in low frequency/high frequency between the two groups of patients （P＞ 0.05）. The other indices of heart rate variability （SDNN， pNN50， very low frequency， low frequency， high frequency， total power） in patients with bipolar depression were lower than those in patients with unipolar depression， and the difference was statistically significant （P＜ 0.001）. Logistic regression showed that except for age， gender， and low frequency/high frequency， all other variable indicators had statistically significant differences in their respective differentiation models （OR=0.959-1.097， 95%CI： 0.937-1.166， P＜ 0.05）， and the difference in body mass index among all differentiation models was statistically significant （OR=1.076-1.097， 95%CI：1.010-1.166， P ＜ 0.05）. Conclusions The heart rate variability index may become a psychophysiological biomarker for distinguishing between unipolar depression and bipolar depression.