In the present work we present the initial stages and first results of the development of a site-specific intelligent plant protection system for bean cultivation. The aim of the system is to be used as a decision tool for rational management of the major key pests of bean crop and with the view to reduce plant protection costs and to mitigate side effects and negative impact on the environment. Currently, we have established a telemetric meteorological network which consists of seven meteorological stations which are distributed in the major bean cultivation area of Greece and over the border area of the Prespes National Park. Through climate sensors the network delivers real time weather data to a cloud server to be further used to establish real time risk thresholds of pest occurrence during the bean growth season. Additionally, the system is passing a stage of pest models development and risk thresholds using in field phenology data along with weather data and Growing Degree-days. Particularly, for the development and validation of the pest risk thresholds we have established experimental fields. The risk thresholds include the pests: Helicoverpa armigera, Thrips sp. and Tetranychus urticae. Pest phenology is observed in four experimental fields which consist of two conventional as well as two organic fields serving as control. The current results will contribute to a better exploitation of the established meteorological station network and their real time microclimate data to provide important information on the rational use of pesticides.