An AI agent's capacity to recall past interactions, learned information, and current objectives is fundamental to its performance in complex, multi-step tasks. This chapter examines how prompt engineering techniques facilitate memory management within agentic systems. We will address methods for working with short-term memory, such as information within an agent's context window, and discuss approaches for long-term information retrieval from knowledge bases. You will learn to construct prompts that guide agents in summarizing data, querying external knowledge stores, maintaining informational consistency, and sustaining task focus across prolonged interactions. The chapter also includes practical considerations for prompting agents that process evolving information streams.