Water management is an essential task for all crop production. However, it is difficult to determine short term plant water needs, as the plant does not exhibit readily detectable indicators of stress until well beyond optimum water conditions. The plant utilizes water in several critical ways, such as, to maintain turgidity, for nutrient uptake, and in photosynthesis, all of which vary in proportion to the environmental conditions. The traditional procedure for supplying plant water needs is to provide a storage of water within the root zone which becomes an immediately available source. Soil or soilless mixes for potted plants have moisture contents which can range from field capacity after watering, to high soil water tensions associated with the onset of wilting conditions of the plant. Alternatively, the root zone volume and storage capacity has for some crops been significantly reduced, requiring a more "on-demand" water feeding schedule. Hydroponic crop production systems, for example, ebb and flood, where there is little or no buffer within the root zone for water, require automated watering schemes which are extremely dependable, reasonably accurate, and uniform in distribution for production of quality crops. In theory, the plant transpires water and thus requires replenishment at rates which are dependent on the plant microclimate (leaf temperature, solar radiation, air humidity, wind speed), as well as, the plant age, morphology, health, and the ease at which the water is available within the root zone. Water requirement can be determined in either of two ways: (1) correlated to plant and its environment with physical or mathematical models, or (2) measured directly with an electronic transducer, such as with stem "sap flow" device. Each has been applied to selected plants species with reasonable success, but generally maintaining some margin of safety, through a water storage buffer within the root zone. The most practical application of each procedure has been in minimizing over-watering and minimizing plant stress while utilizing traditional irrigation techniques. The "speaking plant" approach can provide new opportunities for application of water and nutrients, and ultimately for control of plant growth, but such procedures require that one must "listen" to the speaking plant. Tjie challenge is to focus on the development of sensors to interpret the plant indicators, and then to respond to them within a control system. Machine vision which utilizes the spectral features (by reflectance), or morphological features (physical shape or dynamic growth response) of the plant is one relatively new option for determining the real-time plant condition. Real-time sensors that directly monitor the plant and its water requirements, which are non-intrusive, non-invasive, reliably calibrated, and integrated within a microclimate control system will be necessary for the ultimate success of such systems. In this paper, the plant water requirements, the delivery systems and the potential of automated monitoring of plant water status within integrated crop production systems will be discussed.