Abstract: This study proposed a method to reconstruct the full-field responses of beam and plate-type structures from measurements recorded by a limited number of displacement or by acceleration sensors using spatial compressive sensing theory. Based on the fact that sine series can quickly approximate structural mode shapes, general dictionaries for both beam and plate-type structures are constructed. The number of sensors is determined by analyzing the sine series terms needed to approximate the modal shapes, while their installation positions are determined by the Latin Hypercube Sampling algorithm. The sparse representation vector is solved using the Orthogonal Matching Pursuit algorithm and employed for full-field response recovery. The effects of various factors on the method proposed are assessed through a parametric study, followed by validation through a laboratory test. The presented examples indicate that the method proposed is a widely applicable approach for reconstructing the full-field response of structures, which is expected to significantly improve the efficiency of structural vibration tests and to enhance the accuracy of vibration analysis.