Purpose The purpose of this paper is to contribute to increased knowledge of destructive leadership in crisis management. The specific research questions are: (1) What types of destructive leadership behaviours can be identified in leaders in crisis management? and (2) Why are these behaviours considered destructive in this context? Design/methodology/approach About 21 informants involved in crisis management at regional, local and operational levels in Sweden were interviewed. They were selected since they had recently been involved in severe accidents and/or crises (e.g. terror attacks, forest fires). A grounded theory analysis of interview data yielded two core variables: destructive leadership behaviours, and appraisal: interpretation of leader behaviour. Findings The study identified seven different destructive leadership behaviours: four task-related and three relationship-related. Task-related behaviours primarily led to negative consequences for the task/crisis. Relationship-related behaviours have negative consequences for subordinates' job satisfaction, well-being and/or sense of meaningfulness. The paper relates the identified behaviours to existing leadership ideals within crisis management and discusses behaviours that appear to be unique for the crisis management context. Practical implications The paper highlights the fact that great crisis managers are not always good at managing relationships, which may have negative implications for crisis management in the long term. Originality/value Destructive leadership is a research field that is rapidly expanding. However, there is a lack of knowledge concerning the nature of destructive leadership behaviours and what makes an individual appraise a leader as destructive in crisis management.