Imagine ordering a package and having it delivered by a drone that understands your text messages. That's the vision behind a groundbreaking new system called LLM-DaaS, which harnesses the power of large language models (LLMs) to revolutionize drone delivery services. Current drone delivery systems rely on structured data, which can be a barrier for users who naturally express their needs in everyday language. LLM-DaaS bridges this gap by interpreting free-form text requests like "Deliver a 3kg package to my office at Node 14 from my home at Node 7" and translating them into the precise instructions drones need. This research explores how LLMs like Phi-3.5 and LLaMA were trained to understand these requests, extracting key details such as pickup and drop-off points, package weight, and even delivery time windows. The results are impressive: after fine-tuning, these LLMs achieved accuracy scores near 99% in understanding and structuring user requests. But the innovation doesn't stop there. LLM-DaaS goes beyond simply understanding requests. It intelligently selects the best drone for the job based on factors like battery life, payload capacity, and even real-time weather conditions. If a storm rolls in, the system dynamically adjusts the drone's route using algorithms like A* and Dijkstra's, ensuring the package reaches its destination safely and efficiently. While the system shows incredible promise, challenges remain. Future research will focus on optimizing pathfinding algorithms for even more complex delivery scenarios, managing larger drone fleets, and improving cost-effectiveness. LLM-DaaS is a glimpse into the future of automated delivery, demonstrating how combining the power of LLMs with intelligent systems can transform how we interact with technology and receive goods and services.