By virtue of the combined merits of optical microscopy and flow cytometry, imaging flow cytometry is a powerful tool for rapid, high-content analysis of single cells in large heterogeneous populations. However, its efficiency (defined by the ratio of the number of clearly imaged cells to the total cell population) is not high (typically 50–80%), due to out-of-focus image blurring caused by imperfect fluidic focusing of cells, a common drawback that not only reduces the number of cell images useable for high-content analysis but also increases the probability of false events and missed rare cells. To address this challenge and expand the efficacy of imaging flow cytometry, here, we propose and demonstrate intelligent deblurring of out-of-focus cell images in imaging flow cytometry. Specifically, by using our machine learning algorithms, we show an 11% increase in variance and a 95% increase in first-order gradient summation of cell images taken with an optofluidic time-stretch microscope. Without strict hardware requirements, our intelligent de-blurring method provides a promising solution to the out-of-focus blurring problem of imaging flow cytometers and holds promise for significantly improving their performance.