Background: China''s ''Internet Plus'' nursing services, which are Uber-style home care services with an ''online application, offline service'' approach, have been evolving over the past five years. Registered nurses'' preference for these Uber-style Internet Plus nursing services are crucial for improving human resource management and service efficiency, yet research in this area remains scarce. Objective: This study aimed to explore registered nurses'' preferences for Uber-style Internet Plus nursing services and provide optimization recommendations from a supply-side perspective. Design: A cross-sectional study utilising a discrete choice experiment. Setting(s): Two public tertiary hospitals located in Tianjin, China, which have implemented Internet Plus nursing services. Participants: 211 registered nurses who participated in Internet Plus nursing services. Methods: The survey was conducted anonymously using an online survey platform. Respondents were presented with choices between two alternatives, based on five key attributes: income, safety and security, patient and family cooperation, commute time, and service type. Mixed logit models estimated the stated preferences for attributes. Relative importance scores, willingness-to-pay estimates, and simulations of service-type uptake rates were calculated. Subgroup analysis and seemingly unrelated regression estimation were performed to examine heterogeneity in preferences. Results: A total of 3202 choice observations were generated. When sorted by the strength of preference, the five attributes related to registered nurses'' choice of Uber-style Internet Plus nursing services, measured by their relative importance scores, are as follows: safety and security (30.89 %), income (27.41 %), patient and family cooperation (18.47 %), service type (11.96 %), and commuting time (11.27 %). Elevating safety and security from low to high levels has the same utility as a 31.81 % increase in monthly income, equivalent to 2586.14 yuan. Subgroup analysis showed that senior nurses place more value on safety and security than junior nurses (beta = 1.421 vs.beta = 0.725; P = 0.011), and unmarried nurses had a stronger preference for family and caregiver cooperation (beta = 1.105 vs.beta = 0.314; P = 0.023). Conclusions: The strength and heterogeneity of registered nurses'' preferences should be highlighted in the dispatch algorithms model of Uber-style Internet Plus nursing services, thereby enhancing the efficiency and humanity of Uber-style Internet Plus nursing services. Tweetable abstract: : Registered nurses prioritise safety and security, acknowledging heterogeneous preferences in Uber-style Internet Plus nursing services.