In this paper, a direction-of-arrival (DoA) estimation method based on compressive sensing is proposed for an electronically steerable parasitic array radiator (ESPAR) antenna, which uses only a single radio frequency (RF) chain, and is thereby suited for application in compact wireless terminals. Unlike a conventional multi-active antenna array, signals impinging on parasitic elements in an ESPAR array cannot be processed, and only the output of the sole active element can be processed. In this context, for an ESPAR array, a sparse representation of the DoA estimation problem is formulated by first using an overcomplete dictionary composed of samples from the array manifold and then projecting them onto a set of directional beampatterns. The projection matrix is designed to divide the angle space of the receive antenna array into sectors which are accessed via their corresponding sector beampatterns formed on a time division basis. The sparse signal spectrum is reconstructed by the l1-SVD (singular value decomposition) method [1], where the sparsity is enforced by the l1-norm penalty. Simulation results are presented to demonstrate the efficiency of the proposed method.