Objectives The aim of this study is to estimate the risk of hip fracture after first‐ever stroke, using a nationwide population‐base data set and a retrospective cohort design. Materials and methods The cohort study involved 18,413 patients surviving a first‐ever stroke during the 12‐year period from 1997 to 2008. Another 18,413 control subjects were randomly selected with adjustment for age, gender and enrolled year. Stroke type, duration between stroke and hip fracture, six comorbidities and five categories of medication prior to hip fracture were investigated. Results This study found that 788 (4.3%) subjects in the study group suffered from hip fracture, with a 4.2 years median time frame (interquartile range = 1.8–7.1). In the control group, 492 subjects (2.7%) suffered from hip fracture during a 4.8 years median time frame (interquartile range = 2.0–7.5). The relative risk of hip fracture for stroke was increased in the first four years (1.4–2.4) and gradually declined to the level of the general population. Cox regression analysis showed osteoporosis‐related factors, including ageing, female and antidepressants, significantly increased hip fracture risk (hazard ratios 1.89, 1.57, 1.92). Conclusions These findings imply that osteoporosis may play a major role in the occurrence of hip fracture in the first four years after a first‐ever stroke. Early intervention to prevent bone loss should be regarded as an important part in stroke management, especially in older females, and should be sustained for four years at least. The benefit of antidepressants in stroke patients should be weighed against the increased risk of hip fracture.