Background <p>Despite the global increase in annual incidence of hematological malignancies, the associations between early-life exposures and adult hematological malignancies were not well understood.</p> Methods <p>By conducting a prospective cohort study in the UK biobank, hazard ratios (HR) and 95% confidence intervals (CI) for hematological malignancies based on early-life exposures were estimated with Cox regression models. We further conducted a meta-analysis combining our study with previous cohort studies, and a two-sample Mendelian randomization (MR) analysis to evaluate the potential causal association.</p> Results <p>Smoking initiation &lt; 14&#xa0;years old, a larger body size and taller in height at age 10, were associated with an increased risk of hematological malignancies in adulthood, with corresponding HRs 1.17 (95% CI = 1.03; 1.32), 1.16 (95%CI = 1.06; 1.27) and 1.24 (95%CI = 1.14; 1.36), respectively. The associations were mediated by adulthood BMI and height, but not smoking status. Meta-analysis found no significant association for maternal smoking around birth, multiple births and breastfed as a baby. In addition, the MR analysis suggested a potential causal effect of height at age 10 and age at smoking initiation on leukemia (<i>p</i> &lt; 0.05).</p> Conclusions <p>Taller height at age 10 and younger age at smoking initiation were associated with an increased risk of hematological malignancies in adulthood, especially for leukemia. These findings emphasize the importance of comprehensive tobacco control among adolescents and inclusion of these factors in the assessment for hematological malignancies risk.</p>