Learning Management System (LMS) is widely used in higher education. Researchers have proposed methods to analyze the relations among learning objects (i.e., re-sources/activities) of a course in the LMS and then construct the graph structure for the learning objects. Student’s learning behaviour in the LMS can be represented and analysed in graph, i.e., the Learning Object Graph (LOG). With the LOGs represent different students’ learning behaviours, plug-in is designed to cluster students into groups based on their learning behaviours. Such method requires the relations among learning objects can be identified and measured accurate and properly. This research explains how the LORD (Learning Object Relation Discovery) Moodle plug-in measures the similarity between two learning objects, with the help of WordNet and Natural Language Processing, according to their content in English, French and Hindi to create a more reasonable and objective Learning Object Graph (LOG) that can be used to represent students’ sequential behaviours among learning objects.