Background Attention‐deficit/hyperactivity disorder (ADHD) aggregates in families. To date, the strength, pattern, and characteristics of the familial aggregation have not been thoroughly assessed in a population‐based family sample. Methods In this cohort study, we identified relative pairs of twins, full and half‐siblings, and full and half cousins from 1,656,943 unique individuals born in Sweden between 1985 and 2006. The relatives of index persons were followed from their third birthday to 31 December 2009 for ADHD diagnosis. Birth year adjusted hazard ratio (HR), that is, the rate of ADHD in relatives of ADHD‐affected index persons compared with the rate of ADHD in relatives of unaffected index persons, was estimated in the different types of relatives using Cox proportional hazards model. Results During the follow‐up, 31,865 individuals were diagnosed with ADHD (male to female ratio was 3.7). The birth year adjusted HRs were as follows: 70.45 for monozygotic twins; 8.44 for dizygotic twins; 8.27 for full siblings; 2.86 for maternal half‐siblings; 2.31 for paternal half‐siblings; 2.24 for full cousins; 1.47 for half cousins. Maternal half‐siblings had significantly higher HR than in paternal half‐siblings. The HR did not seem to be affected by index person's sex. Full siblings of index persons with ADHD diagnosis present at age 18 or older had a higher rate of ADHD (HR: 11.49) than full siblings of index persons with ADHD diagnosis only before age 18 (HR: 4.68). Conclusions Familial aggregation of ADHD increases with increasing genetic relatedness. The familial aggregation is driven by not only genetic factors but also a small amount of shared environmental factors. Persistence of ADHD into adulthood indexes stronger familial aggregation of ADHD.