Abstract. Approximate Petri nets (AP-nets) can be used for the knowledge representation and approximate reasoning. The AP-net model is defined on the basis of the rough set approach, fuzzy Petri nets and coloured Petri nets. One of the main advantages of AP-net model is a possibility to present the reachability set of a given AP-net by means of an occurrence graph. Such graphs can serve, among others, for analyzing and evaluating an approximate reasoning realized by using AP-net model. The main contribution of the paper is to present the algorithms for construction and analysis of occurrence graphs for the AP-nets, especially in the context of searching for the best decision and finding the shortest distance in order to compute such decision. This approach can be applied to the design and analysis of the formal models for expert systems, control systems, communication systems, etc.