An efficient high-capacity quantum secret sharing scheme is proposed following some ideas in quantum dense coding with two-photon entanglement. The message sender, Alice prepares and measures the two-photon entangled states, and the two agents, Bob and Charlie code their information on their photons with four local unitary operations, which makes this scheme more convenient for the agents than others. This scheme has a high intrinsic efficiency for qubits and a high capacity. [3,4,5,6,7], quantum dense coding [8], quantum secure direct communication (QSDC) [9,10], and so on. A surprising property of an entangled quantum system is its nonlocality. Two parts of the quantum system cannot be considered to be independent even if they are far apart, and the singleparticle measurements on these two parts cannot give all the information about the state of the whole quantum system. Quantum nonlocality has been embodied in the process of quantum teleportation [2], an important quantum technique. Ekert exploited the nonlocality feature to design a QKD protocol [3] in 1991, and Bennett, Brassard and Mermin (BBM92) [4] simplified its error rate analysis process in 1992. Also, quantum nonlocality has been used to transmit a secret message directly [9,10].Secret sharing is a classical cryptographic scheme [11,12,13] in which a boss, say Alice suspects that one of her two remote agents, say Bob and Charlie, may be dishonest but she does not know who the dishonest one is. She believes that the honest agent can prevent the dishonest one from destroying her benefits if they act in concert. For the security of her message M A , Alice splits it into two pieces M B and M C , and sends them to Bob and Charlie, respectively. The two agents can read out the message M A = M B ⊕ M C only when they cooperate. As a classical signal can be copied perfectly, it is impossible to create a private key with classical physics. When quantum mechanics enters the field of information, the story is changed. Quantum secret sharing (QSS) is the generalization of classical secret sharing into quantum scenario and has progressed quickly in recent years [14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29 One of the main goals of QSS, similar to QKD, is to distribute the private keys among the three participants, or more generally, many participants securely [14,15,16,17,18,19,20,21,22,23,24,25,26]. An original QSS scheme was proposed by Hillery, Bužek and Berthiaume [14] in 1999, which is called HBB99 hereafter. In the HBB99 scheme [14], the secret sharing is accomplished by using a three-photon entangled GreenbergerHorne-Zeilinger (GHZ) state. Each participant holds a photon from a GHZ state, and chooses randomly one measuring-basis (MB) from the X-MB and the Y -MB to measure their photons independently, similar to BBM92 QKD scheme [4]. Subsequently Karlsson, Koashi and Imoto (KKI) put forward another QSS scheme [15] with a two-photon polarization-entangled state. The photons are polarized along the z or x directions, and the two agents measure their photon...