Objectives. Kaixinsan (KXS), a traditional Chinese medicine formula, has been demonstrated to be effective in the treatment of depression. The present study applied a network pharmacology approach to dig out the new targets and mechanism of action of KXS and the active compounds in the treatment of depression. Methods. A network pharmacology approach based on public databases including ADME (absorption, distribution, metabolism, and excretion) evaluation, targets prediction, construction of networks, and molecule docking was used and validated the predicted new antioxidant targets and mechanisms in vitro. Based on an in vitro experiment, we verified the AKT1/Nrf2 pathway related to the thioredoxin (Trx) antioxidant mechanism. Results. The present study sorted 31 pharmacologically active components (kaempferol, ginsenoside rh2, ginsenoside rh4, stigmasterol, etc.) through the ADME algorithm from KXS. 136 potential molecular targets (AKT1, TNF, IL-1b, JUN, ESR1, NOS3, etc.) were predicted, of which there were 69 targets clearly related to depression. By compound-depression targets (C-DTs) network constructed, and protein-protein interaction networks (PPI) and KEGG pathway enrichment analyzed, we identified active compounds mediating depression-related targets to exert synergism on the predictive AKT1/Nrf2 pathway related to thioredoxin (Trx) antioxidant mechanism and other inflammation-related signaling pathways as well as neurotransmitter related signaling pathways. In the H2O2 induced SH-SY5Y cell damage model, this showed kaempferol and ginsenoside rh2 could enhance the activity of the Trx system by upregulation of AKT1 to activate Nrf2 in vitro. Conclusions. Taken together, by comprehensive systems pharmacology approach analysis, we found that KXS and its active compounds might exhibit antioxidant effects by stimulating the AKT1/Nrf2 pathway in the treatment of depression, which might shed new light on innovative therapeutic tactics for the new aspects for depression in traditional Chinese medicine in future studies.