The northern Arabian Sea is highly productive in terms of abundant and diverse fish fauna (≈1,000 species) due to nutrient-rich waters. A total of 303 specimens were collected from the coastal waters of Pakistan, representing 100 genera, 61 families, and 15 orders identifiable to 120 morphological species and 18 others that could not be morphologically assigned to any known taxa. However, the COI analysis via neighbor-joining (NJ) tree, Bayesian inference (BI), and Maximum-likelihood (ML) trees and Automatic Barcode Gap Discovery (ABGD) produced nearly similar topologies and revealed 138 single-species-representing phylogenetic clusters including 18 highly supported previously unknown species clusters showing deep intraspecific divergences from the closely related morphospecies. The calculated average Kimura 2-parameter (K2P) genetic distances were 0.25%, 6.34%, 17.12%, and 26.8% within the species, genera, families, and orders, respectively. We analyzed species from eight groups (clupeiforms, sciaenids, flatfishes, spiny-flathead, lizardfish, sillaginids, grunts, stargazers) through BI and ML and found cryptic genetic diversity in these groups. Barcoding gap analysis (BGA) demonstrated that the barcode gaps were present for 39 of 70 species having multiple specimens (50.7%). Our study provides the first-ever addition of COI sequences for five species (Cociella punctata Cuvier, 1829; Epinephelus stoliczkae Day, 1875; Pomadasys aheneus McKay & Randall, 1995; Scorpaenopsis possi Randall & Eschmeyer, 2002; and Sillago panhwari Panhwar et al., 2017 in Panhwar et al. 2018) into the GenBank database. The study concluded that DNA barcoding is an efficient approach in species identification, supplementing the morphological identification of the known forms and highlighting the cryptic diversity.