Background: Smoking is strongly associated with higher risk of psoriasis in several observational studies; however, whether this association is causal or can be explained by confounding or reverse causation is not fully understood. Randomized controlled trials are the gold standard when examining causality; however, when this method is not feasible, the Mendelian randomization design is an alternative. Herein genetic variants can be used as robust proxies for modifiable exposures and thereby avoiding confounding and reverse causation. In this study, we hypothesized that smoking is an independent and causal risk factor for psoriasis and tested this using a Mendelian randomization design. Methods: We used data from the Copenhagen General Population Study including 105,912 individuals with full information on lifestyle factors, biochemistry, and genotype data. In total, 1,240 cases of moderate to severe psoriasis were included to investigate the association between smoking and psoriasis. To assess causality of the association, we used the genetic variant CHRNA3 rs1051730, where the T-allele is strongly associated with high lifelong cumulative smoking, as a proxy for smoking. Results: In observational analyses, the multivariable adjusted hazard ratio of developing moderate to severe psoriasis was 1.64 (95% confidence interval: 1.35-2.00) in ever smokers with ≤ 20 pack-years and 2.23 (1.82-2.73) in ever smokers with > 20 pack-years compared to never smokers. In genetic analyses, the odds ratio of developing moderate to severe psoriasis was 1.05 (0.95-1.16) per CHRNA3 rs10511730 T-allele in ever smokers. Conclusion: Smoking was an independent risk factor for moderate to severe psoriasis in observational analyses. However, using a genetic variant as a robust proxy for smoking, we did not find this association to be causal.