Security of data is an issue that is of significant interest. In this paper, we propose a new compressive sensing-based data encryption system that can represent the original signal with far fewer samples than the conventional Nyquist sampling-based system. Compressive sensing could also be treated as an encryption algorithm with good secrecy. As an application example, we apply it to sense-through-wall ultra-wideband (UWB) noise radar that requires enormous storage space and high security. Interestingly, a random Gaussian matrix is sufficient to capture the information of UWB noise radar signal; no knowledge of UWB signal is required in advance. Simulation results indicate only one-third of the original samples are needed to perfectly recover UWB noise radar signal, and compressive sensing provides good secrecy as an encryption algorithm. It is impossible to retrieve the original message without the entire sensing matrix. Copyright © 2013 John Wiley & Sons, Ltd.