Sociological research shows that friendship and partner choice tend to reveal a bias toward social similarity. These relations are ruled by the so called “proximity principle” which states that the more similar two individuals are, the more likely they will become friends. However, proximity, similarity or friend- ship are concepts with blurred edges and grades of membership (acquaintances, friends, couples). Therefore, in order to model the friendship dynamics we work on an Agent-Based Model that already manages the social relationships, together with demographics and evolutionary crossover. To introduce these theoretical concepts we decided to fuzzify the system, explaining the process in detail. Thus, we end up with fuzzy sets and operations, a fuzzy friendship relationship, and a logistic function for its evolution.