As an emergent-architecture, mo-bile edge computing shifts cloud service to the edge of networks. It can satisfy several desir-able characteristics for IoT systems. To reduce communication pressure from IoT devices, data aggregation is a good candidate. Howev-er, data processing in MEC may suffer from many challenges, such as unverifiability of ag-gregated data, privacy-violation and fault-tol-erance. To address these challenges, we pro-pose PVF-DA: privacy-preserving, verifiable and fault-tolerant data aggregation in MEC based on aggregator-oblivious encryption and zero-knowledge-proof. The proposed scheme can not only provide privacy protection of the reported data, but also resist the collusion be-tween MEC server and corrupted IoT devices. Furthermore, the proposed scheme has two outstanding features: verifiability and strong fault-tolerance. Verifiability can make IoT device to verify whether the reported sensing data is correctly aggregated. Strong fault-tol-erance makes the aggregator to compute an aggregate even if one or several IoTs fail to report their data. Finally, the detailed security proofs are shown that the proposed scheme can achieve security and privacy-preservation properties in MEC.