Reactive oxygen species (ROS) are involved in the regulation of many physiological processes. However, overproduction of ROS under various cellular stresses results in cell death and organ injury and thus contributes to a broad spectrum of diseases and pathological conditions. The existence of different cellular sources for ROS and the distinct properties of individual ROS (their reactivity, lifetime, etc.) require adequate detection methods. We therefore compared different models of cellular stress and various ROS-sensitive dyes-2',7'-dichlorodihydrofluorescein diacetate (DCF-DA), MitoSOX™, and MitoTracker® red CM-H(2)XRos-using a confocal fluorescent imaging approach, which has the advantage of not only detecting but also of localizing intracellular sources for ROS. Confocal acquisition of DCF-DA fluorescence can be combined with ROS detection by the mitochondria-specific probes MitoSOX™ and MitoTracker® red CM-H(2)XRos. Specificity was controlled using various antioxidants such as Trolox and N-acetylcysteine. Using different fluorescent ROS-sensitive probes, we detected higher ROS production equally under cell starvation (IL-3 or serum depletion), hypoxia-reoxygenation, or treatment of cells with prooxidants. The detected increase in ROS was approximately threefold in IL-3-depleted 32D cells, approximately 3.5-fold in serum-deprived NIH cells, and 2.5-fold to threefold in hypoxic HL-1 cells, and these findings agree well with previously published spectrofluorometric measurements. In some cases, electron spin resonance (ESR) spectroscopy was used for the validation of results from confocal fluorescent imaging. Our data show that confocal fluorescent imaging and ESR data are in good agreement. Under cellular stress, mitochondrial ROS are released into the cytoplasm and may participate in many processes, but they do not escape from the cell.