The study was carried out to investigate the body shape variation of the Mystus gulio collected from three different salinity gradient habitats (high-salinity brackish water, low-salinity brackish water, and freshwater) and two different sources (wild and hatchery) using linear morphometric distances, landmark-based truss-networking and geometric morphometrics. A total of 436 individuals of M. gulio were collected from the high-salinity (8–15 ppt), low-salinity (2–6 ppt), and freshwater bodies (0 ppt). For wild versus hatchery sources body shape variation, 430 individuals of M. gulio were collected from various coastal rivers and hatcheries. M. gulio from different salinity gradient habitats and sources exhibited negligible sexual dimorphism. Principal component analysis (PCA) indicated that high and low salinity populations appeared to form overlapping clusters with the freshwater populations, but other multivariate analyses discriminated both populations from brackishwater habitats from the freshwater counterparts. The geometric morphometrics displayed that the body shape variation of M. gulio from salinity gradient habitats was mainly visible in the width of the dorsal-ventral part, the snout shape, the tail shape, the head region, and the eye diameter. PCA demonstrated that multivariate spaces of both the wild and captive populations overlapped each other, although many individuals of the wild population were discriminated from the culture populations. Our results showed that truss networking and geometric morphometric methods provide consistent outcomes for body shape discrimination in the M. gulio populations.