<p>Mobile crowdsourcing has emerged as a powerful paradigm where multiple task requesters submit diverse tasks to be executed by heterogeneous IoT devices acting as task executors in strategic setting. In such strategic environments, both task requesters and executors may misreport their private preferences to obtain favorable allocations, which threatens fairness, efficiency, and participation. To address these challenges, this paper introduces TMTMMMC, a truthful and Pareto-optimal two-sided matching mechanism tailored for preference-aware task allocation in IoT-based mobile crowdsourcing. The proposed framework models each requester’s heterogeneous tasks and each executor’s distinct capacity using the notion of virtual task requesters and virtual task executors. Both sides submit private preference lists over their feasible counterparts. A novel mechanism is designed that ensures each virtual requester is matched to the best possible virtual executor in a way that is truthful, computationally efficient, and Pareto optimal. Theoretical analysis confirms that the mechanism guarantees strategy-proofness, Pareto optimality, and efficiency, while extensive simulations demonstrate superior performance over benchmark random mechanisms in terms of truthfulness, the proportion of agents obtaining their first preferences, average preference positions, and runtime scalability.</p>