Previous observational studies showed associations between obstructive sleep apnea (OSA) and depression and other negative moods. However, the causality has not been determined. Single nucleotide polymorphisms were identified as instrumental variables by screening from genome-wide association studies. Bidirectional two-sample mendelian randomization (MR) was applied to assess the potential causal relationship between OSA and depression, subjective well-being, negative moods. Inverse variance weighting (IVW) method and weight median were chosen as the main methods to estimate possible causal effects. MR-Egger, MR pleiotropy residual sum and outlier and leave-one-out analysis methods, were used as sensitivity analysis methods to ensure robust results. MR analyses indicated significantly causal association of OSA on depression (OR = 1.22, P = .010) and major depressive disorder (OR = 1.02, P = .006). Furthermore, genetically predicted OSA was negatively associated with subjective well-being (β