The increasing prevalence of non-tuberculous mycobacterium (NTM) infections alongside tuberculosis (TB) underscores a pressing public health challenge. Yet, the mechanisms governing their infection within the lung remain poorly understood. Here, we integrate metagenomic sequencing, metabolomic sequencing, machine learning classifiers, SparCC, and MetOrigin methods to profile bronchoalveolar lavage fluid (BALF) samples from NTM/TB patients. Our aim is to unravel the intricate interplay between lung microbial communities and NTM/ infections. Our investigation reveals a discernible reduction in the compositional diversity of the lung microbiota and a diminished degree of mutual interaction concomitant with NTM/TB infections. Notably, NTM patients exhibit a distinct microbial community characterized by marked specialization and notable enrichment of and , driving pronounced niche specialization for NTM infection. Simultaneously, these microbial shifts significantly disrupt tryptophan metabolism in NTM infection, leading to an elevation of kynurenine. , , , and have been implicated in the metabolic pathways associated with the conversion of indole to tryptophan via tryptophan synthase within NTM patients. Additionally, indoleamine-2,3-dioxygenase converts tryptophan into kynurenine, fostering an immunosuppressive milieu during NTM infection. This strategic modulation supports microbial persistence, enabling evasion from immune surveillance and perpetuating a protracted state of NTM infection. The elucidation of these nuanced microbial and metabolic dynamics provides a profound understanding of the intricate processes underlying NTM and TB infections, offering potential avenues for therapeutic intervention and management.