This article supplements the array of methods in urban policy discourse analysis by applying textometry to a corpus of planning documents. Textometry is a systematic computer-assisted analysis of textual data. When applied to large corpora, the method can reveal contrasts that cannot readily be detected by non-instrumented human reading. It seems promising therefore to use textometry to analyse urban planning documents which, in France as in other countries, are known for being at first sight pretty indistinguishable from each other. By analysing dozens of documents at a time, this method increases the scale of analysis, enabling us to identify massive transitions in discourse over time or major contrasts among discourses emanating from specific groups of actors. When combined with classical methods (i.e. interviews, qualitative archival research), textometry seems to be effective at identifying new ways of understanding urban policy discourses. To exemplify the potential of such a method, the analysis developed here draws on a corpus of 36 French urban transport plans (plans de déplacements urbains) for the period 2000–2015. Our results display marked contrasts in discourses, mainly as a result of changes over time. Paradoxically, for most cities between 2000 and 2010, the discourses become increasingly abstract in their content while highlighting a proactive attitude and providing specific information on the document implementation process and a precise list of institutional partners.