Mobile Crowdsensing (MCS) leverages distributed mobile devices to collect and analyze data, enabling large-scale and real-time data acquisition from diverse environments. A key challenge in MCS is task allocation, which impacts data quality, resource utilization, and user engagement. Existing methods often struggle with low allocation efficiency, inaccurate task assignments, and a lack of privacy protection, which reduces user participation. This paper addresses these issues by proposing a privacy-preserving multi-task allocation approach based on a fuzzy inference system. The method employs differential privacy to prevent privacy leakage. It integrates a fuzzy controller capable of dynamically adjusting main parameters during the operation of a genetic algorithm, providing a more robust and flexible solution. To validate the effectiveness of our proposed approach, we conduct experiments on real-world datasets. Results demonstrate that the proposed method achieves higher efficiency and effectively protects user privacy, providing a feasible task allocation solution for mobile crowdsensing.