Abstract: As personal data have been widely collected and shared by various application platforms, the risks of privacy intrusion and data breaches have been substantially increased. To prevent malicious access to or misuse of private data, explicit user consent and controllability over personal data are required to be ensured by data collectors. The mainstream approach of binding consent to data-processing purposes, however, still exhibits several shortcomings: purpose-based consent reasoning is deprived of clear semantics, the association between purposes and platform business logic is overlooked, and effective mechanisms for conflict resolution are lacking, making it difficult for valid consent to be obtained. To address these issues, a consent-based access control model is proposed, in which consent permissions are incorporated into access-decision processes and all operations that process personal data are regarded as actions performed for specific purposes. Semantically, a “purpose” is defined as a precondition or an effect of an action, and action consent is mapped to user authorization, thereby establishing an explicit semantic linkage among purpose, action, and consent. Following the principles of least privilege and deny-by-default, asymmetric consent-inheritance rules over purpose hierarchies are proposed and a consent-computation method is provided. The model is formalized using TLA+, and several key security properties are verified through model checking. Experimental results indicate that, compared with representative methods, the proposed consent computation more effectively resolves consent conflicts and yields superior outcomes.