Recent advancements in sensor development and network technologies have increased the demand for lightweight machine learning techniques that operate efficiently in computational resource-constrained environments, such as edge AI. Among these methods, reservoir computing has attracted attention as a form of lightweight machine learning. Reservoir computing enables fast learning owing to its low computational cost; however, achieving high accuracy requires proper reservoir design, with long-lasting chaotic dynamics generally preferred. In this study, we investigated whether the Sugarscape Game of Life (GoL) can function as a reservoir. The Sugarscape GoL, a model incorporating elements from Sugarscape, is expected to exhibit a range of chaotic dynamics depending on parameter settings. Experiments using random initial configurations demonstrated that the transient step count (step count needed to reach equilibrium) was significantly increased compared with the conventional GoL, and chaotic dynamics persisted for extended durations. In addition, varying the parameters altered the transient step count, producing diverse chaotic dynamics. Clarifying the characteristics of the Sugarscape GoL and applying them to reservoir computing can contribute to technologies that enhance daily convenience and industrial productivity in an IoT-driven society.