Objective To investigate the use of psychiatric drugs in hospitalized patients with depressive disorder， and to provide references for clinical treatment. Methods Using the Beijing-Tianjin- Hebei big data platform， the data of eligible patients with depressive disorder admitted to Beijing Anding Hospital from 2015 to 2019 were retrospectively analyzed， and the use of psychotropic drugs in patients with early-onset or late-onset depressive disorder was compared. Results A total of 6 043 inpatients were enrolled， including 1 820（ 30.12%） patients with early-onset depressive disorder and 4 223（ 69.88%） patients with late-onset depressive disorder. The main classification of antidepressants used in early-onset patients were SSRIs（ 61.54%， n=1 120）， SNRIs（ 21.26%， n=387） and NaSSA（ 7.25%， n=132） respectively， including sertraline（ 25.07%， n=456）， escitalopram（ 21.98%， n=400） and venlafaxine（ 13.68%， n=249）. The use rate of antipsychotics was 68.08%（ n=1 239）， including quetiapine（ 29.67%， n=540）， aripiprazole（ 15.93%， n=290） and olanzapine（ 10.11%， n=184）； the use rate of mood stabilizers was 33.85%（ n=616）， including lithium carbonate（ 18.08%， n=329）， valproate（ 15.44%， n=281） and lamotrigine（ 0.33%， n=6）. The main classification of antidepressants used in late-onset patients were SSRIs（ 61.59%， n=2 601）， SNRIs（ 27.97%， n=1 181） and NaSSA（ 19.37%， n=818）， including escitalopram（ 27.73%， n=1 171）， sertraline（ 9.91%， n=841） and mirtazapine（ 19.37%， n=818）； the use rate of antipsychotics was 59.10%（ n=2 496）， including quetiapine（ 30.59%， n=1 292）， olanzapine（ 12.48%， n=527 cases） and aripiprazole（ 5.23%， n=221）； the use rate of mood stabilizers was 8.69%（ n=367）， including lithium carbonate（ 3.84%， n=162）， valproate（ 4.62%， n=195）， lamotrigine（ 0.21%， n=9）. The proportion of early-onset patients without antidepressants was higher than that of late-onset patients［ 12.03%（ 219/1 820） vs. 3.43%（ 145/4 223）； χ2=166.153， P ＜ 0.001］， and the proportion of combination with antidepressants was lower than that in late-onset patients［ 11.65%（ 212/ 1 820） vs. 20.03%（ 846/4 223）； χ2=61.913， P＜ 0.001］. There were statistically significant differences in the use rates of sertraline， escitalopram， fluvoxamine， fluoxetine， paroxetine， mirtazapine， duloxetine， lithium carbonate， valproate， aripiprazole， amisulpride， clozapine and olanzapine between the two groups（ all P＜0.05）. Patients with early-onset depressive disorder were more likely to use SSRIs， and were more likely combined with antipsychotics and mood stabilizers. Conclusions In the real world， there are great differences between the drugs used for patients with early-onset depressive disorder and those with late-onset depressive disorder， which should be paid more attention to in clinical treatment.