Understanding the heterogeneous preferences of individuals for disaster insurance attributes is critical for product improvement and policy design. In an era of global environmental change, the Qinghai-Tibet Plateau is a hotspot of natural hazards. Improving the capability of rural housing disaster insurance to foster local residents’ disaster resilience is of great significance but remains under addressed. We used a discrete choice experiment approach to provide the first estimates of rural residents’ preferences for rural housing disaster insurance attributes in central and western Tibet. We estimated residents’ preferences and willingness-to-pay for the sum insured, subsidy rate, insured object, and perils covered. The potential impacts of increasing the sum insured, expanding the insured object, and lowering subsidy rates were evaluated. Our results suggest that residents prefer products with a high sum insured, high subsidy rate, and a complete list of insured objects. Residents who have experienced specific hazards tend to prefer the corresponding perils covered. Females and residents who have a closer social network are more likely to purchase insurance. Product improvement and policy simulation results suggest that, while lowering the subsidy rate, increasing the sum insured and expanding the insured object could promote participation and improve residents’ welfare. Our results could improve the understanding of the preferences of households in remote regions and support policy implementations.