Objective To describe and compare patients' and nurses' preferences for the implementation of bedside handover. Design Discrete choice experiment describing handover choices using six characteristics: whether the patient is invited to participate; whether a family member/carer/friend is invited; the number of nurses present; the level of patient involvement; the information content; and privacy. Setting Two Australian hospitals. Participants Adult patients (n=401) and nurses (n=200) recruited from medical wards. Main outcome measures Mean importance scores for handover characteristics estimated using mixed multinomial logit regression of the choice data. Results Both patient and nurse participants preferred handover at the bedside rather than elsewhere (P<.05). Being invited to participate, supporting strong two‐way communication, having a family member/carer/friend present and having two nurses rather than the nursing team present were most important for patients. Patients being invited to participate and supporting strong two‐way communication were most important for nurses. However, contrary to patient preferences, having a family member/carer/friend present was not considered important by nurses. Further, while patients expressed a weak preference to have sensitive information handed over quietly at the bedside, nurses expressed a relatively strong preference for handover of sensitive information verbally away from the bedside. Conclusions All participants strongly support handover at the bedside and want patients to participate although patient and nurse preferences for various aspects of bedside handover differ. An understanding of these preferences is expected to support recommendations for improving the patient hospital experience and the consistent implementation of bedside handover as a safety initiative.