Mobile financial services (MFS) expand financial inclusion in emerging economies but expose users to phishing, SIM-swap fraud, and social engineering risks inadequately addressed by static awareness programs. This study develops and empirically validates a Dynamic Cybersecurity Awareness (DCA) framework integrating the Technology Acceptance Model and Protection Motivation Theory to jointly explain adoption and protective behaviour. Using a stratified survey of 1,159 Kenyan MFS users and Structural Equation Modelling, the study tests relationships among governance, training, trust, threat awareness, self-efficacy, and secure digital behaviour. Results show that trust (β=0.286) and threat awareness (β=0.257) exert the strongest effects on secure behaviour, while training and governance act as enabling mechanisms through self-efficacy and trust. The validated DCA framework establishes adaptive feedback loops that reinforce protective practices, offering policymakers and providers a scalable, evidence-based framework for reducing fraud and strengthening resilient digital financial ecosystems.