Availability of solar radiation in the urban environment is determined to a great extent by urban geometry, namely how densely built-up an area is and how the given built volume is distributed spatially within the site. This paper explores relationships between urban geometry and solar availability on building façades and at the pedestrian level, with implications for buildings’ passive potential and outdoor thermal comfort, respectively. The study was based on the morphological and solar analysis of 24 urban forms of London, covering a wide range of built density values found across the city. Two aspects of solar availability were investigated at the neighbourhood scale, through statistical analysis: (i) the relationships between urban geometry variables and solar availability indicators in different time periods, and (ii) the seasonal solar performance of urban forms’ façades and ground. Apart from the strong, negative effect of density, the analysis revealed that solar availability on ground and façades is significantly affected by urban layout. Mean outdoor distance, site coverage, directionality and complexity were the most influential for the solar performance of open spaces; while building façades were mostly affected by complexity, standard deviation of building height and directionality. However, direct solar irradiance on ground and façades was found to be influenced by different variables in January and July, which is attributed to the different solar altitude angles. Related to that, urban forms have been identified that present higher irradiance values in January and lower in June when compared to others. Considering temperate climates, these examples highlight the potential for enhancing the seasonal solar performance of existing and future urban developments. Finally, the seasonal effect on solar availability appears to be much more pronounced for ground with its mean direct irradiance value increasing on average by a factor 15, from January to July, while for façades the increase is only by a factor 2.6.