Roguelike games are renowned for their procedural generation of immersive and unpredictable game worlds. A key element within these games is the inclusion of secret rooms, which often contain valuable rewards or challenging encounters. Determining the content of these secret rooms in a manner that strikes a balance between randomness and player engagement is crucial for creating a captivating gameplay experience. This abstract presents a novel approach for determining secret room content in roguelike games using a Fuzzy Markov Process (FMP). While traditional Markov processes have been widely employed for procedural content generation, their deterministic nature can lead to repetitive gameplay patterns. By integrating fuzzy logic into the process, the FMP enables the generation of secret room content that is dynamic, organic, and less predictable. The FMP incorporates multiple factors, including player progression, game difficulty, and thematic coherence, as fuzzy sets and employs fuzzy transition matrices. Through experimental evaluation and player feedback, the effectiveness and player satisfaction of the FMP-based secret room generation system can be assessed. This research contributes to advancing the state-of-the-art in roguelike game design by providing a robust framework for generating intriguing secret room content.