We conducted a two-sample Mendelian randomization study to determine the association of smoking initiation with seven psychiatric disorders. We used 353 independent single-nucleotide polymorphisms associated with cigarette smoking initiation as instrumental variables at genome-wide significance threshold (p < 5 × 10-8) from a recent genome-wide association study in 1,232,091 European-origin participants. Summary-level data for seven psychiatric disorders, including anxiety, bipolar disorder, insomnia, major depressive disorder, posttraumatic stress disorder, suicide attempts, and schizophrenia, was obtained from large genetic consortia and genome-wide association studies. The odds ratios of genetically predicted smoking initiation were 1.96 for suicide attempts (95% CI 1.70, 2.27; p = 4.5 × 10-20), 1.69 for post-traumatic stress disorder (95% CI 1.32, 2.16; p = 2.5 × 10-5), 1.54 for schizophrenia (95% CI 1.35, 1.75; p = 1.6 × 10-10), 1.41 for bipolar disorder (95% CI 1.25, 1.59; p = 1.8 × 10-8), 1.38 for major depressive disorder (95% CI 1.31, 1.45; p = 2.3 × 10-38), 1.20 for insomnia (95% CI 1.14, 1.25; p = 6.0 × 10-14) and 1.17 for anxiety (95% CI 0.98, 1.40; p = 0.086). Results of sensitivity analyses were consistent and no horizontal pleiotropy was detected in MR-Egger analysis. However, the associations with suicide attempts, schizophrenia, bipolar disorder, and anxiety might be related to possible reverse causality or weak instrument bias. This study found that cigarette smoking was causally associated with increased risks of a number of psychiatric disorders. The causal effects of smoking on suicide attempts, schizophrenia, bipolar disorder and anxiety needs further research.