Objective- Whether tobacco smoking causally affects white and red blood cells and thrombocyte counts is unknown. Using a Mendelian randomization approach, we tested the hypothesis that smoking causes increases in these blood cell indices. Approach and Results- We included 104 607 white Danes aged 20 to 100 years from the Copenhagen General Population Study with information on blood cell indices, smoking habits, and CHRNA3 (alpha 3 nicotinic cholinergic receptor) rs1051730 genotype, where the T allele causes higher tobacco consumption; 41 759 were former smokers and 17 852 current smokers. In multivariable adjusted observational analyses and compared with never smokers, white blood cells were associated with up to 19% increases, thrombocytes with up to 4.7% increases, and red blood cell indices with up to 2.3% increases in former and current smokers. All associations were dose dependent, with tobacco consumption but for white blood cells and thrombocytes also dependent on smoking cessation time in former smokers; highest increases were for <1-year smoking cessation and lowest increases for >10-year smoking cessation. In age- and sex-adjusted genetic analyses, percent differences per T allele increase in current smokers were 1.15% (95% CI, 0.61%-1.68%) for leukocytes, 1.07% (0.38%-1.76%) for neutrophils, 1.34% (0.66%-2.02%) for lymphocytes, 1.50% (0.83%-2.18%) for monocytes, -0.60% (-1.91% to 0.74%) for eosinophils, 0.17% (-0.94% to 1.29%) for basophils, 0.38% (-0.17% to 0.93%) for thrombocytes, 0.04% (-0.14% to 0.23%) for erythrocytes, 0.34% (0.17% to 0.50%) for hematocrit, 0.26% (0.09% to 0.43%) for hemoglobin, and 0.29% (0.18% to 0.41%) for mean corpuscular volume. Conclusions- Smoking causes increased blood leukocytes, neutrophils, lymphocytes, and monocytes, as well as increased hematocrit, hemoglobin, and mean corpuscular volume. The observational smoking relationships were long term for white blood cells and short term for red blood cell indices.