Irrigation decisions based on traditional soil moisture sensing often leads to uncertainty regarding the true amount of water available to the plant. Plant based sensing of water stress decreases this uncertainty. In specialty crops grown in California’s Central Valley, precision deficit irrigation based on plant water stress could be used to decrease water use and increase water use efficiency by supplying the necessary quantity of water only when it is needed by the plant. However, there is a lack of a clear decision support system to implement stress-based precision irrigation on a management unit basis. Management zones were developed using an unsupervised fuzzy classification technique, where zone divisions were based on the soil characteristics of digital elevation, shallow electrical conductivity, sand, silt content and plant characteristics (leaf temperature and canopy cover). Management zones were most influenced by digital elevation. In both almonds and grapes, two zones were identified and two treatments were implemented. A leaf monitor was used to proximally sense the leaf temperature and microclimatic variables (relative humidity, air temperature, wind speed, and incident radiation) and compute a crop water stress index in almond and grape crops. Temperature differences between leaf surface and air were used to obtain stress indices. A wireless mesh network system was used to interface 14 leaf monitors in the almond plot while 10 leaf monitors interfaced to hubs equipped with cellular modems were used in the grape plot. In both cases data were transmitted to the web where they could be accessed in real-time to guide irrigation decisions. Plant water stress was estimated with respect to a well-watered and a simulated dry control tree. Variable rate irrigation was applied in almonds according to crop water stress levels in each management zone. Plant stress was evaluated in grapes in eight groups of five vines over a period of 10 days, where four groups were recovering through irrigation after experiencing severe stress and the other four groups were stressed by not applying water to the vines. In grapes, MCWSI and DSWP were found to be linearly related with a coefficient of determination value of 0.82.  In almonds, a strong correlation resulted from comparison of CWSI and DSWP, obtained for each treatment of each zone; this yielded a second order polynomial relationship with a coefficient of multiple determination value of 0.79. In almonds, preliminary results indicated that Zone 1 required only 70% water compared to the grower treatment while Zone 2 required about 90% water compared to the grower treatment. These results suggest this method has the potential for increased water savings and increased water use efficiency.