It is necessary to study the combined toxicity of an herbicide and its safener because the two are often used in combination. S-metolachlor and its safener benoxacor have been detected in aquatic environments and can individually damage the oxidative stress system in zebrafish embryos (Danio rerio). However, only their separate toxicity in zebrafish (Danio rerio) embryo development has been reported. This study assessed the combined toxicity of benoxacor and S-metolachlor in zebrafish embryo development, including acute toxicity, developmental toxicity, oxidative damage, and cell apoptosis. The 96-h LC50 values were higher in mixtures of benoxacor and S-metolachlor than in benoxacor alone. The treatments included S-metolachlor, Mix-1 (0.1 mg/L benoxacor + 0.1 mg/L S-metolachlor), Mix-2 (0.1 mg/L benoxacor + 0.3 mg/L S-metolachlor) and benoxacor alone. Embryos exposed to Mix-1 and Mix-2 had lower developmental toxicities, superoxide dismutase (SOD) activity, osx and cat expression levels than those exposed to benoxacor alone. Moreover, glutathione S-transferase (GST), catalase (CAT), glutathione reductase (GR), glutathione peroxidase (GPx) activities, and the expressions of tbx16, nrf2, bcl2, and caspase9 were higher in the mixtures than in the benoxacor group. High-throughput RNA sequencing revealed that benoxacor had a greater effect on gene regulation than Mix-1 and Mix-2. The malformation rate, different enrichment gene numbers, and gene expression levels of hatched embryos were higher in Mix-1 than in Mix-2. The results indicate that a mixture of S-metolachlor and benoxacor has antagonistic effects in the early stage of embryo development. The mixtures can break the reactive oxygen species balance, causing abnormal cell apoptosis and developmental malformation in embryos. Besides investigating the combined toxicity of benoxacor and S-metolachlor in zebrafish embryo development, this study provides a risk assessment basis for a herbicide combined with its safener.