ABSTRACT Background Tense production is often impaired in people with aphasia (PWA). Interestingly, the literature suggests that not all tenses are affected in the same way, although the pattern of dissociation is still debated in the literature. The heterogeneity of tense production deficits in PWA might partly explain this lack of consensus in the literature. While this heterogeneity has been demonstrated by descriptive statistics or double dissociations in several studies, no study has used cluster analyses to highlight the main patterns of tense dissociation. Furthermore, the factors underlying this heterogeneity have been little explored. Aims Our study aimed to use cluster analyses on tense production performance to (1) identify the main patterns of dissociation between tenses in PWA and (2) explore the clinical (aphasia type and severity) and cognitive (executive disorders and temporality) factors that may underlie these patterns. Methods & procedures Twenty-one French-speaking participants with fluent and non-fluent aphasia completed a verb inflection production task, as well as language and cognitive tests. Cluster analyses were performed on the differences in performance between the tenses on the verb inflection production task. Generalized linear mixed models and nonparametric statistics were used to analyze the effect of tense and its interaction with the clusters, and to compare the clusters on the clinical and cognitive variables. Outcomes & results Cluster analyses revealed two main clusters, one with worse performances in the past and future than in the present (PWA_1 – 70% of the PWA), and the other with worse performances in the past and present than in the future (PWA_2 – 30%). The type and severity of aphasia, and cognitive test performance did not differ between the two clusters. On the other hand, the PWA_2 cluster had a longer time post-onset than the PWA_1 cluster. Conclusions Our results confirm the heterogeneity of verb inflection deficits in PWA. They suggest that most PWA would have difficulties with past and future tenses, questioning the presence of discourse linking in these two tenses. However, a minority of PWA, who are clinically and cognitively indistinguishable from the others, do not follow this pattern. Adopting the strategy of producing a morphologically simple tense in French – future – could explain the high performance in the future tense and the errors in substituting other tenses with the future in this subgroup. Finally, the clinical and research implications of this heterogeneity, in relation to therapies and the lack of consensus in the literature, are discussed.