Wideband wireless channel is a time dispersive channel and becomes strongly frequency-selective. However, in most cases, the channel is composed of a few dominant taps and a large part of taps is approximately zero or zero. They are often called sparse multi-path channels (MPC). Conventional linear MPC methods, such as the least squares (LS), do not exploit the sparsity of MPC. In general, accurate sparse MPC estimator can be obtained by solving a LASSO problem even in the presence of noise. In this paper, a novel CS-based sparse MPC method by using Dantzig selector (DS) is introduced. This method exploits a channel's sparsity to reduce the number of training sequence and, hence, increase spectral efficiency when compared to existed methods with computer simulations.