The competitiveness of present and future district heating systems can be at risk when residential and service sector heat demands are expected to decrease in the future. In this study, the future competitiveness of district heating has been examined by an in depth analysis of the distribution capital cost at various city characteristics, city sizes, and heat demands. Hereby, this study explores an important market condition often neglected or badly recognised in traditional comparisons between centralised and decentralised heat supply. By a new theoretical approach, the traditional and empirical expression for linear heat density is transformed into an analytical expression that allows modelling of future distribution capital cost levels also in areas where no district heating exists today. The independent variables in this new analytical expression are population density, specific building space, specific heat demand and effective width. Model input data has primarily been collected from national and European statistical sources on heat use, city populations, city districts and residential living areas. Study objects were 83 cities in Belgium, Germany, France, and the Netherlands. The average heat market share for district heat within these cities was 21% during 2006. The main conclusion is that the future estimated capital costs for district heat distribution in the study cities are rather low, since the cities are very dense. At the current situation, a market share of 60% can be reached with a marginal distribution capital cost of only 2.1 (sic)/GJ, corresponding to an average distribution capital cost of 1.6 (sic)/GJ. The most favourable conditions appear in large cities and in inner city areas. In the future, there is a lower risk for reduced competitiveness due to reduced heat demands in these areas, since the increased distribution capital cost is low compared to the typical prices of district heat and competing heat supply. However, district heating will lose competitiveness in low heat density areas. Hence, reduced heat demands in high heat density areas are not a general barrier for district heating in the future. (C) 2010 Elsevier Ltd. All rights reserved.