Artificial neural networks and expert systems slowly go into every aspect of our life, replacing us in very difficult and time-consuming activities. These systems are most common where specialized knowledge is required to solve the complex problems. One of the most effective methods of decision support used in such systems are decision trees. They are characterized by transparency, simplicity and ability to solve the complex classification problems. Currently on the market, both among the window-based applications and websites, there is no environment that facilitates the creation of expert decision support systems, based on the method of decision trees. Therefore, the aim of the paper was to design and develop scattered computer system which would allow the user to generate decision tables and trees for agricultural advisory.