This chapter delves into the methodologies of prompt engineering to enhance the functionality and precision of Large Language Models (LLMs). It begins by highlighting the role of prompt engineering in aligning LLM outputs with specific user needs and intentions. Then categorizes prompting techniques into several types, each designed to improve how LLMs understand and respond to tasks: Zero-shot prompting; One-shot prompting; Few-shot prompting; and Chain-of-thought prompting. Chain-of-thought guide the LLM through a logical sequence to solve tasks requiring detailed analysis. By employing semiotics-the study of signs and symbols and their use or interpretation-the prompts can be designed to navigate the model through structured reasoning paths, addressing both the syntactic and semantic layers of language processing. This perspective helps in crafting prompts that act as mediating agents between the user and the LLM, aligning the machine’s outputs with human semantic intentions and enhancing the interaction quality.