In order to automatize crop farming and its processes, a number of technological and other problems have to be solved. Agricultural field robots are in our vision to fulfill operations in fields. Robots involve number of technological challenges in order to be functional and reliable, but also systems controlling these robots are to be developed. In this paper automatic crop farming is the vision, and decision making models and operational planning is discussed. Study is carried out with simulation environment capable of simulating crop growth, weather, automatic crop farm storages and sheds, robots and stationary systems, and systems controlling the robots. The driving force for simulation is spatial variation of soil, crop growth potential and variable solar radiation; and this sets challenges for automatic decision making. This study concentrates on fictional barley crop farm in Southern Finland. The operations to be done are conventional: harrowing and seeding together with local fertilizing, weed control, supplementary fertilizing, crop disease control and harvesting. Automatic decision making models are studied, and developed for each operation. Operational planning involves question that in which order fields are operated. Decision making for seeding including harrowing and fertilizing is considered as one operation, and the developed model relies on stationary online moisture and temperature measurements from field; and decision for seeding is done when a certain threshold is exceeded. For weed and crop disease control operations a fuzzy logic is utilized and it works using normalized measurements. The most challenging operations from a point of decision making are supplementary fertilizing and harvesting, as they have a strong influence over incoming weather, and therefore the decision making model requires probabilistic and prediction based approaches. The developed decision making model for supplementary is based on simulated crop growth model over the rest of farming season and varying weather scenarios.