Neurotransmitter uptake by sodium-coupled monoamine transporters of the NSS family is required for termination of synaptic transmission. Transport is tightly regulated by protein-protein interactions involving the small cytoplasmic segments at the amino (N-) and carboxy (C-) terminal ends of the transporter. Although structures of homologs provide information about the transmembrane regions of these transporters, the structural arrangement of the terminal domains remains largely unknown. Here, we combined molecular modelling, biochemical and biophysical approaches in an iterative manner to investigate the structure of the 82-residue N-terminal and 30-residue C-terminal domains of human serotonin transporter (SERT). Several secondary structures were predicted in these domains and structural models were built using the Rosetta fragment-based methodology. 1-dimensional (1)H NMR and circular dichroism (CD) spectroscopy were consistent with the presence of helical elements in the isolated SERT N-terminal domain. Moreover, introducing helix-breaking residues within those elements altered the fluorescence resonance energy transfer (FRET) signal between terminal CFP and YFP tags in full-length SERT, consistent with the notion that the fold of the terminal domains is relatively well defined. Full-length models of SERT were generated that are consistent with these and published experimental data. The resultant models predict confined loci for the terminal domains whose separation increased during the transport-related conformational cycle, as predicted by structures of homologs and by the 'rocking bundle' hypothesis, and consistent with spectroscopic measurements. The models also suggest the nature of binding to regulatory interaction partners. This study provides a structural context for functional and regulatory mechanisms involving SERT terminal domains.