In recent years, people-centric sensing networks have attracted much research effort. To date, there are still some significant security and privacy challenges in people-centric sensing networks. In this paper, we focus on the private data sharing and protection in people-centric sensing networks. First, we formalize the network model with relay nodes which improves the data forwarding efficiency of networks. Second, we propose a novel Attribute based Private data sharing protocol in People-centric sensing networks (APP). Relying on the technology of ciphertext policy attribute based encryption, our APP protocol can protect the privacy and integrity with efficient approaches of authentication, encryption, transmission and decryption. Also, we propose an associative data indexing scheme to improve the private data sharing performance. Finally, we discuss the performance evaluation of APP protocol in detail and find that it can achieve much better efficiency.