The present dissertation intends to determine a structure for vegetable oils production and supply of biodiesel plants – from crops seeds distribution to oil extraction – regarding agricultural, logistics and industrial aspects. For that purpose, a computational model considering a stochastic approach was designed and implemented as a tool for decision making. A two level solution is provided: oil extraction plants are detailed for their size, technology, number of units and location, as well as crops production is geographically distributed in order to respond to the biodiesel plant annual demand. The designed model was then tested in a case study considering castor seeds plantation in the northern part of Minas Gerais state for supply of a biodiesel plant located in the city of Montes Claros. Basic solution was known for a deterministic scenario and was followed by sensibility analyses for some selected parameters such as freight costs, investment costs increase and biodiesel plant demand increase. Some other social restrictions were also considered regarding, for example, a minimum number of farmers contracted annualy. Finally some aspects relating to yield variations and probabilities were treated using the stochastic model, and the option for this kind of modeling was then analysed based in two important measurements: EVPI and VSS.