There is a sound theoretical base and intuitive expectation that urban office markets are too segmented to be accurately described by unitary market models. Nevertheless, previous studies of office market dynamics tend to concentrate on either national or metropolitan markets and assume a unitary market in equilibrium. In spite of the fact that the value of office property attributes may vary across urban submarkets.Office submarkets can be understood as comprising offices that, although not exactly alike in their combination of characteristics, are considered a reasonable substitute for each other. In general, two types of segmentation can be distinguished; spatial segmentation and structural segmentation. Spatial segmentation refers to geographical delineated submarkets due to dissimilar location features. Structural segmentation relates to market segmentation due to building dissimilarities such as building size, building quality and year of development. The study examines the spatial and structural segmentation of the Amsterdam office market. Amsterdam, the study area of this research, has a dispersed office location pattern. Amsterdam is by far the largest office centre in The Netherlands and has an office stock of approximately 6.7 million square meters. Real estate agents identify several distinct office locations; spatial submarkets. Furthermore, distinct building stock segments, relating to building size and building period, are distinguished. In this paper, the Amsterdam submarket structure is examined in a hedonic framework, over the period 1995 – 2011. Prior defined submarkets, as perceived by real estate agents, are tested on consistency and submarkets are statistically derived. For this purpose a standard city-wide hedonic model is formulated and compared with models that take various submarket delineation schemes in consideration. The derived utility for building and location attributes in each model is compared and provides an indication of the submarket structure of Amsterdam.