As mentioned at the beginning of the book, narrative generation has been decomposed into at least two stages, producing a plot, and then generating text from the plot. This chapter first looks at planning mechanisms that guide the formulation of a story. Historical story generation approaches focused on characters and their goals, creating plans relying on hand-crafted commonsense knowledge to achieve those goals, and generating texts from it. Other approaches used case-based reasoning, leveraging pre-existing story fragments and then generalizing them as needed. Included among these are systems that made use of narrative functions. To control the narrative better, systems started representing narrative goals in addition to character goals. The chapter also describes interactive narrative, where the audience or players can shape the narrative, and may take on the role of a character, and where models of the audience can be used to control outcomes. Next, the chapter discusses temporal generation, which leverages many of the notions related to time discussed in the previous chapter. The chapter then arrives at contemporary neural approaches to story generation. Here the chapter discusses both the use of user-generated plots as prompts, as well as system-generated plots. An important trend is the use of repositories of commonsense knowledge, which allows one to capture some aspects of character goals and causality that were the hallmark of historical approaches. The chapter examines the challenges and approaches to evaluation, before providing an updated NarrativeML for story generation. The chapter also assesses the status of story generation systems today. Finally, it stresses the importance of attempting to treat the code that produced the story output as a first-class object of literary analysis, in keeping with the approach of the field of Critical Code studies.