This paper defines a strong security model for multi-recipient signcryption scheme in the random oracle model, which is semantically secure against adaptive chosen-ciphertext insider attack, existentially forgeable chosen-message attack and anonymous chosen-ciphertext insider attack. By analyzing, we point out that Li-Hu-Liu Multi-Recipient Signcryption Scheme is insecure against the adaptive and anonymous chosen-ciphertext insider attacks. Later, we present an improved scheme, which is proven to be secure in our newly defined strong security model.