This chapter evaluates the Brain-Inspired memory model in the context of Thai word segmentation, a task complicated by the absence of explicit word boundaries. It introduces two brain-inspired methods, THDICTSDR and THSDR, both of which leverage Sparse Distributed Representations (SDRs) to encode lexical patterns and contextual cues. Unlike traditional or deep learning approaches, these methods demonstrate an ability to manage ambiguity and noise without relying on large training datasets or complex architectures. The chapter reviews prior segmentation techniques and highlights their limitations, particularly in handling unknown words. Empirical evaluation on benchmark datasets shows that the proposed SDR-based models achieve competitive accuracy and exhibit strong robustness to noise. These findings support the broader applicability of brain-inspired models in natural language processing and suggest promising directions for further research beyond conventional methodologies.