Wildland fire management has moved beyond a singular focus on suppression, calling for wildfire management for ecological benefit where no critical human assets are at risk. Processes causing direct effects and indirect, long-term ecosystem changes are complex and multidimensional. Robust risk-assessment tools are required that account for highly variable effects on multiple values-at-risk and balance competing objectives, to support decision making. Providing wildland fire managers with risk-analysis tools requires a broad scientific foundation in fire behaviour and effects prediction as well as high quality computer-based tools and associated databases. We outline a wildfire risk-assessment approach, highlight recent developments in fire effects science and associated research needs, and recommend developing a comprehensive plan for integrated advances in wildfire occurrence, behaviour and effects research leading to improved decision support tools for wildland fire managers. We find that the current state of development in fire behaviour and effects science imposes severe limits on the development of risk-assessment technology. In turn, the development of technology has been largely disconnected from the research enterprise, resulting in a confusing array of ad hoc tools that only partially meet decision-support needs for fuel and fire management. We make the case for defining a common risk-based analytic framework for fire-effects assessment across the range of fire-management activities and developing a research function to support the framework.