Urbanization is currently one of the most widespread disturbances urgently requiring empirical data regarding its effects on coastal ecosystems. The aim of this study was to compare patterns of variability in populations of the Mediterranean mussel, Mytilus galloprovincialis, between urban and non-urban intertidal rocky shores, over a temporal scale of 12 months and multiple spatial scales (from cm to 10 s of km). For this, variance components associated with percentage cover, spat and total density, condition index, shell length and clump thickness of mussels were compared. Different patterns emerged depending on the response variable and the spatial and temporal scale. There was in general, a higher variability in urban than in non-urban shores, particularly for shell length, spat and total density that can be interpretated as a first stage of degradation, before noticing changes in mean values of these variables. Moreover, the most relevant scales of variability of total and spat density changed with urbanization (10 s of km in urban; 10 s of cm/m in non-urban). Results highlight the need for adopting proper management plans that should include the relevant spatial and temporal scales of variability; otherwise, they will fail in ameliorating urbanization effects on intertidal ecosystems.