Graphs are a general and powerful data structure that can model a large variety of concepts. Finding the similarity between two graphs and defining a measure of similarity between a set of graphs are important topics that can have various applications: malware detection, model checking, computer vision, and many more. In this work, the similarity between graphs will be evaluated computing the Maximum Common Subgraph (MCS) both using an exhaustive algorithm, which compares two graphs, and both using an algorithm inspired by a general search algorithm which can instead compute the approximate similarity between a set of graphs. To compute the similarity between graphs, it is necessary to define the equivalence between vertices: given two graphs A and B, a vertex x belonging to A is equivalent to a vertex y belonging to B, if two conditions are met: they are of the same type (which is defined by the specific problem), and if some of x’s parents and/or children vertices belong to some equivalence class, also y must have matching parents and/or children in the same equivalence classes.