Aims This study aimed to determine the extent to which nurses report assessing evidence-based falls risk factors and implementing targeted prevention for medical and surgical patients in China. Design This study was a national online survey. Methods The respondents were registered nurses working in medical and surgical units in 662 Chinese hospitals. The data concerning the falls risk factor assessments and targeted interventions implemented by nurses were collected online by the Nursing Management Committee of the Chinese Nursing Association in China in 2019. Results In total, 68 527 valid questionnaires were returned (95.0%). In medical and surgical units, nurses were most likely to report assessing balance, mobility and strength (81.6%) and orthostatic hypotension (76.4%) in falls patients and least likely to report assessing continence (61.3%) and feet and footwear (55.8%). Ensuring the use of appropriate footwear (79.3%) and managing syncope, dizziness and vertigo (73.8%) were the most common multiple interventions, while managing postural hypotension (48.8%) and cognitive impairment (48.4%) was the least common. Nine falls risk factors with clearly matched multifactorial interventions were identified in medical and surgical units (68.2%-97.1%). Conclusions The implementation of multifactorial interventions in medical and surgical wards is inconsistent as reported by nurses in medical and surgical wards. Throughout China, nurses are generally concerned about falls risk factors and prevention for their patients; however, limited attention has been focused on continence, feet and footwear assessment and the management of cognitive impairment. Evidence-based falls prevention should be further tailored to the specific risk factors of each patient. Impact Best practice guidelines for falls prevention in hospitals have been developed and published, and it is important for nurses to use these guidelines to guide practice. Our findings identify that in routine care, healthcare providers and hospitals can prevent falls.