Rivest-Shamir-Adleman　(RSA)　is　one　of　the　widely　deployed　public-key　algorithms.　Yet,　its　decryption　facet　is　very　time　consuming　for　resource-constrained　Internet-of-Thing　(IoT)　devices,　as　it　is　based　on　the　modular　exponentiation　of　a　large　number.　Although　several　variants　of　RSA　have　been　designed　to　accelerate　decryption,　the　outcomes　have　been　far　from　satisfactory.　Therefore,　it　is　of　imminent　importance　to　investigate　how　to　securely　outsource　RSA　decryption　to　computational　powerful　parties　as　an　alternative　solution.　In　this　article,　we　introduce　the　first　efficient　and　secure　outsourcing　scheme　for　RSA　decryption　in　IoT.　Though　RSA　decryption　is　achieved　via　modular　exponentiation,　existing　secure　outsourcing　schemes　for　modular　exponentiation　either　assume　the　modulus　to　be　prime　and　are　not　applicable　to　RSA　or　incur　massive　computation　costs　and　are　heavy　laden　in　practice.　To　address　these　issues,　we　have　designed　our　scheme　based　on　the　Chinese　remainder　theorem　(CRT).　In　our　scheme,　the　private　keys　(including　the　exponent　and　the　modulus)　and　the　plaintext　are　concealed　concurrently,　and　the　proposed　scheme　is　highly　efficient　for　both　client　and　cloud.　In　addition,　our　scheme　enables　the　client　to　detect　any　misbehavior　of　the　cloud　server　with　a　probability　of　99.17%.　To　validate　the　effectiveness　of　our　proposed　scheme,　we　provide　rigorous　proofs　of　security　and　verifiability,　as　well　as　efficiency　analysis.　The　effectiveness　and　efficiency　of　our　scheme　are　further　confirmed　based　on　experimental　results.　©　2014　IEEE.