Abstract: Confidentiality technologies play a pivotal role in driving the continuous advancement of semantic communications. Herein， first， confidentiality technologies in traditional communication， including cryptography based encryption， physical layer key generation， artificial noise， beamforming， and covert communication， are reviewed. Next， existing research on confidential semantic communication—cryptography-encrypted， covert， physical layer-encrypted， and application layer-encrypted—are systematically summarized， and challenges in this area are analyzed. Subsequently， a super-resolution reconstruction-based end-to-end encryption transmission scheme for image semantic communication is proposed； this includes semantic extraction， joint source-channel coding， encryption and decryption， and residual dense network-based super-resolution reconstruction modules. This scheme avoids the high complexity introduced by the complex encryption algorithm and effectively resists eavesdroppers. Additionally， the scheme deploys a super-resolution reconstruction module at the receiver to recover the image details lost due to encryption and decryption， and channel noise， and then improves the visual quality of the image. By comparing the transmission performance under different modulations， signal-to-noise ratio， and scaling factors， the simulation results in additive white Gaussian noise channel verify that the scheme can balance communication efficiency and reconstructed image quality under the premise of ensuring transmission security. Finally， some future research directions in semantic communication， including lightweight design of confidential semantic communication， multimodal encrypted semantic communication， semantic importance-based encrypted semantic communication， semantic privacy protection in multi eavesdropping threat environment， and intention oriented secure communication between machines， are prospected.