A new approach for Confocal Microscopy (CM) based on the framework of compressive sensing is developed. In the proposed approach, a random set of pinholes generated by a spatial light modulator (SLM) is used as illumination and detection mask to eliminate out-of-focus and scattered light at the detector. Unlike conventional CM or programmable array microscopy (PAM) that measures pixel intensity in spatial domain directly, in the compressive confocal microscopy (CCM), the measured data emerges from random sets of pinhole illuminated pixels in the specimen that are linearly combined (projected) and measured by a single pixel detector. Compared to conventional imaging systems, the number of measurements needed for nearly perfect reconstruction in CCM is significantly reduced potentially leading to low cost and simpler acquisition system. Three-dimensional image reconstruction schemes are studied that exploit the intra-slice spatial image correlation as well as the inter-correlation existing between images along the axial dimension. Experimental results are shown based on a testbed that uses an XGA DMD for computing the linear projections of illuminated pixels and a photomultiplier tube for obtaining the image projections. The state of each micromirror in the DMD is defined by a compressive sensing measurement ensemble, optimally designed to avoid micromirror’s crosstalk and maximize, at the same time, light efficiency and spatial resolution. CCM works well in both fluorescence and reflection mode. Theoretical analysis, experimental and simulation results are provided to illustrate the features of the systems.