Analogies facilitate understanding across domains, enabling individuals to navigate unfamiliar concepts using what they already know. Yet, crafting effective analogies requires extensive knowledge of both the source and target domains, making it a challenging task for educators—particularly when aiming to support novice learners. Recent advancements in Large Language Models (LLMs) have demonstrated their ability to generate analogies to explain scientific concepts. However, analogy in education must align with students’ prior knowledge and cognitive resources, which shape how analogies are perceived. To address this challenge, we examine the theory and practices of using analogies in education and introduce a system that generates personalized analogies to facilitate learning. Preliminary evaluations show promising outcomes, and we anticipate deeper insights from future human-subject studies.