This paper presents a multi-agent model of knowledge acquisition in group learning inspired by human cognitive and social characteristics. We present a model that represents knowledge acquisition using a weighted graph, where concepts and their relationships are dynamically explored by learning agents. Agents acquire, reinforce, and share knowledge through communication, simulating real-world collaborative learning processes. The model captures the dynamic nature of knowledge acquisition, where individual characteristics, such as interest and learning ability, influence learning and teaching effectiveness. We provide a formal description of the model, detailing how agents interact within a fixed knowledge environment, and how knowledge propagates over time, and the analysis of the example. We also suggested further directions for the development of the model and its potential applications in other domains.