The relationship between accessibility, as a measure of the fixed location of the property, and land value is well recognized. Space Syntax theory was developed as a set of tools to analyze relationships between structures and functions of cities introducing accessibility measures of the urban grid. Even though admitted that location parameters comprise the most influential factors of urban property value, accessibility measures of centrality, based on network (integration and choice) quantifying urban morphology, have gained little attention in land value literature, despite the fact that urban grid morphology has a crucial role in property market structure. Recently, there has been growing interest for spatial statistics which count for spatial aspects of phenomena such as the land market. The frequently used OLS regression adopted for statistical inference on variables influencing a phenomenon is mostly inefficient for comprehension of spatial phenomena mainly due to instabilities caused by spatial autocorrelation and fixed parameter assumption. GWR extends the classic regression model by allowing spatially varying coefficient estimations while also accounting for spatial autocorrelation. The goal of the research presented herein is to estimate which spatial accessibility radius better explains objective land value and to explore local spatial relationships between geometric accessibility and land values using GWR, Space Syntax theory and GIS techniques, in Xanthi city, a medium sized city in Northern Greece. This research has led to inferences concerning the importance of Space Syntax geometric accessibility in the interpretation of land values, with local patterns of accessibility influence emerging.