Sugarcane is an important sugar and potential energy crop, and the complexity of its genome has led to stagnant progress in genome decipherment and hindered the genome-wide analysis of the nucleotide binding site leucine-rich repeat (NLR) receptor until the genome of Saccharum spontaneum was published. From the genome of S. spontaneum, 724 allelic and non-allelic NLRs were identified and classified into five types (N, NL, CN, CNL, and P) according to domain architectures and integrity and at least 35 genes encoded non-canonical domains. The phylogenetic analysis indicated NLRs containing the coiled-coil (CC) domain separated from those without CC in six Poaceae species, including S. spontaneum. The motif analysis determined the characteristics and potential functions of the 137 representative non-allelic NLRs, especially the core motifs contained in the NBS and LRR domains, which indicated that motifs were regularly distributed among clades. Through transcription factor binding site (TFBS) profiles, we predicted that the most important transcription regulator of NLRs in sugarcane were ERF, MIKC_MADS, and C2H2. In addition, based on three sets of transcriptome data from two sugarcane hybrids and one S. spontaneum clone infected by the necrotrophic fungal pathogen Stagonospora tainanensis causing sugarcane leaf blight (SLB), the expression dynamics of NLRs responding to the infection in three sugarcane clones were compared. The different genetic background led to the significant difference of NLRs response to SLB in different sugarcane clones, and we got an inference of the potential mechanism of SLB resistance. These results provided a basic reference and new insights to further study and utilize the NLRs.