As a cost-effective solution to energy conservation, behavior based method focuses on changing people's behavior through normative feedback for energy efficiency. While the application of behavior-based method is promising, the challenge exists to achieve efficiently sustainable behavioral change. Based on multi-period observation of energy behavior at the Joint Base Andrews in Maryland, this paper presents a model-based approach aimed to improve the nationally popular and deep-seated benchmark setting strategy for normative feedback used in home energy reports. The improved approach has its merits of countering the undesirable boomerang effect and enhancing the effectiveness of normative feedback targeting different personalities. By introducing a modified opinion dynamics model, this paper simulates the process of energy behavior change and therefore identifies the driver and elementary rules of behavioral change. In particular, the paper defines various behavioral zones in accordance with people's personality and proposes a new customized energy reporting mechanism that maps normative benchmark to personality trait. The new energy reporting policy has strong industrial implication for promoting behavior-based method towards a sustained energy conservation movement. (C) 2015 Elsevier Ltd. All rights reserved.