NaY zeolite is used as support for optimal dispersed molybdenum-sulfide to study the effect of dispersion on the thiophene HDS. Two molecules Mo(CO)(6) are deposited inside the supercages of NaY by both a static as well as a flow type preparation method. During deposition inside the supercages by the how method 1-1.5 CO ligand per Mo atom is removed. For the formation of Mo-sulfide particles in the zeolite cavities the sulfidation has to be carried out in the absence of water. Once the Mo sulfide particles are formed in the pores in this way, treatment with H2O or NH3 does not change the catalytic activity. In contrast herewith, the activity strongly decreases when the sulfidation is carried out with humidified H2S due to the formation of well-structured MoS2 at the external zeolite surface. The NaY supported molybdenum-sulfide samples are characterized by EXAFS, AAS, XRD and TEM measurements. According to EXAFS, after stepwise sulfidation of Mo(CO)(4.5-5)/NaY up to 773 K small Mo-sulfide species are formed inside the NaY supercages consisting of two Mo atoms with RMo-Mo=3.16 Angstrom (like in MoS2) surrounded by S and denoted by 'Mo(3.16 Angstrom)S-x'. The formation of these dimer species inside the zeolite supercages takes place via the intermediate formation of the dimer 'Mo(2.78 Angstrom)S-x' species with RMo-Mo=2.78 Angstrom (like in MoS3). In all the formed Mo-sulfide species also some additional light atoms are contributing to the Mo radial distribution function. However, these atoms could not be identified in an unambiguous way and, hence, troubled an easy straight forward analysis of the EXAFS data as will be described. It turned out that the 'Mo(2.78 Angstrom)S-x' clusters formed after sulfidation at temperatures lower than 673 K have a considerable higher thiophene HDS activity than the 'Mo(3.16 Angstrom)S-x' ones formed after sulfidation at 773 K. Comparing the activity of the very well dispersed Mo-sulfide particles inside the zeolite supercages with that of MoS2 supported on Al2O3 and TiO2 shows that differences in HDS activity can not only be explained by differences in dispersion. Possibly, clusters of a certain size are necessary or the support affects the electronic structure of the formed MoS2 species.