Persuasive technologies are designed to influence people and induce them to change their attitudes and behaviours, bringing advantages to their users. Behaviour change support systems are at the heart of persuasive technology research. However, the sought benefits cannot be achieved if the systems fail to engage and retain the users. The present study provides a detailed description of a theory-driven effort to empirically (N=314) explain and predict users' continuance intention towards a behaviour change support system for weight loss. Deriving from extant theories, a research model is constructed and tested through partial least-squares (PLS) analysis. In the proffered model, primary task support affects perceived effort and perceived effectiveness. Computer-human dialogue support has strong connections to primary task support, perceived social support and perceived effectiveness. Perceived credibility has a significant relationship to the continuance intention. Social identification has a strong connection to perceived social support, which, in turn, has a significant effect on perceived effectiveness and continuance intention. Finally, perceived effectiveness has a significant impact on use continuance. Investigating the aspects related to the continued use of behaviour change support systems is feasible, as it will guide future implementations of such systems.