There are 936 million daily active users on Facebook who mutually exchange messages, share photos, update their status and similar. Results of previous studies have shown that it is possible to predict user's relationship in real life based on such user interactions. The purpose of this work is to examine the existing model for determining the strength of friendship based on data on the interaction of users on Facebook and additionally determine whether the model can predict the trust between users. Through the developed web application survey was taken on the users of Facebook and a significant amount of data was collected and stored in the database. By analyzing the data, it was concluded that the model can be used to group friends into certain groups depending on the strength of their friendship and that it can predict 5% of friends in which the user has the most trust. The collected data is the basis for future research to determine the boundaries between groups of friends more accurately, and to improve the model.