Numerical reconstruction is a key procedure of digital holography. Classical reconstruciton algorithms are characterized by the twin images, superimposition of in-focus and out-of-focus object, and so on. Application of the new compressive sensing theory to sparsity reconstruction of digital inline holography is investigated. Based on the linear operator of diffraction model, the algorithm for the reconstruction of digital inline holography using compressive sensing is derived. Holograms of particle field from both simulation and micro digital inline holography are reconstructed, and the reconstructed particle field with compressive sensing algorithm is compared with that of convolution algorithm. Results show that the compressive sensing algorithm can improve the quality of reconstructed planes, and can reconstruct the particle field with good accurary with even 25% of the hologram. It also shows potential in focusing performance and robustness to noise.