In smart homes, users remotely manage resource utilization tasks and context-aware services using portable devices and mobile communication technologies. Reliability of automated energy consumption management relies upon context-aware continuous authentications of users in executing time-critical tasks. In particular, the contexts of mobility of users and the critical nature of the task are important. Continuous authentication is a robust technique to ensure validity of the authenticity of users over time. Existing continuous authentication techniques do not use the contextual information and dynamic user behavioral characteristics for authentications. We propose a novel context-aware multi-attribute continuous authentication model for secure energy utilization management in smart homes. We use location and the critical nature of the tasks as the contextual information as supporting information for selecting the authentication attributes. We propose novel location and task profiles as context specification metrics and a novel relative-importance based attribute selection technique based on N-model. The usefulness of the proposed solution is validated using real-world data sets. Furthermore, the reliability of the proposed risk based resource management model is analysed as a constraint model using linear temporal logic. Based on the experimental results, this research provides meaningful insights to use pragmatic approaches with security and reliability assurances for resource management applications in smart homes. (C) 2015 Elsevier Ltd. All rights reserved.