While (N-ethyl perfluorooctanesulfonamido)ethanol (FOSE) -based phosphate diester (diSPAP) has been proposed as a candidate precursor of perfluorooctanesulfonate (PFOS), its potential biotransformation to PFOS has not been verified. Metabolism of diSPAP was investigated in Japanese medaka ( Oryzias latipes ) after exposure in water for 10 days, followed by 10 days of depuration. Branched isomers of diSPAP (B-diSPAP) were preferentially enriched in medaka exposed to diSPAP, with the proportion of branched isomers (BF) ranging from 0.56 to 0.80, which was significantly greater than that in the water to which the medaka were exposed (0.36) (p < 0.001). This enrichment was due primarily to preferential uptake of B-diSPAP. PFOS together with perfluorooctanesulfonamide (PFOSA), N-ethyl perfluorooctanesulfonamide (NEtFOSA), 2-(perfluorooctanesulfonamido)acetic acid (FOSAA), NEtFOSAA, FOSE, and NEtFOSE were detected in medaka exposed to diSPAP, which indicated the potential for biotransformation of diSPAP to PFOS via multiple intermediates. Due to preferential metabolism of branched isomers, FOSAA and PFOSA exhibited greater BF values (>0.5) than those of NEtFOSA, NEtFOSAA, and NEtFOSE (<0.2). Such preferential metabolism of branched isomers along the primary pathway of metabolism and preferential accumulation of B-diSPAP led to enrichment of branched PFOS (B-PFOS) in medaka. Enrichment of B-PFOS was greater for 3-, 4-, and 5-perfluoromethyl PFOS (P3MPFOS, P4MPFOS, and P5MPFOS), for which values of BF were 0.58 ± 0.07, 0.62 ± 0.06, and 0.61 ± 0.05 (day 6), respectively; these values are 5.8-, 7.8-, and 6.4-fold greater than those of technical PFOS. This work provides evidence on the isomer-specific accumulation of PFOS from diSPAP and will be helpful to track indirect sources of PFOS in the future.