Toxic encephalopathy is a central nervous system disorder caused by endogenous or exogenous toxic substances. Granzyme B (GZMB), a key serine protease, plays a crucial role in immune regulation and the progression of various diseases. Abnormal expression of GZMB may contribute to the development and progression of toxic encephalopathy; however, its specific mechanisms remain unclear. This study aims to preliminarily explore the association between granzyme B and toxic encephalopathy through big data. This study retrieved the toxic encephalopathy dataset GSE253309 from the gene expression omnibus database. Differentially expressed genes were identified using the "limma" R package, followed by gene ontology and Kyoto encyclopedia of genes and genomes enrichment analyses. A protein-protein interaction network was constructed using the search tool for the retrieval of interacting genes (STRING) database, key hub genes were identified with Cytoscape software. A total of 994 differentially expressed genes were identified. Among them, GZMB was moderately overexpressed in toxic encephalopathy tissues (log2 fold change = 0.70, P = .00087), suggesting a potential disease association. and was enriched in the TGF-beta signaling pathway. Kyoto encyclopedia of genes and genomes enrichment analysis revealed that GZMB-related genes were involved in several pathways, notably the TGF-beta signaling pathway and apoptosis, both of which are implicated in immune regulation and neuronal injury. Protein-protein interaction network analysis confirmed GZMB as a hub gene potentially contributing to disease progression. Furthermore, β-escin sodium and methylprednisolone may modulate GZMB expression, thereby alleviating neuronal damage and improving outcomes in delayed encephalopathy after carbon monoxide poisoning. Elevated GZMB expression likely contributes to the pathogenesis and progression of toxic encephalopathy through multiple pathways, making it a potential disease biomarker and therapeutic target.