Relatively little is known about how attachment influences children’s relationships to pets or mediates positive and negative interaction outcomes. We carried out in-depth interviews with 27 children, including nine children at high-risk for animal harm and 18 matched controls. We used the Child Attachment Play Assessment (CAPA), a drawing task and self-report measures including the Short Attachment to Pets Scale (SAPS) and Children’s Animal Harm Behaviours (CAHB). We also designed a novel measure, the ‘Pets In Children’s Attachment Stories’ (PICAS), to probe children’s mentalising about pets, caregiving-behaviours, comfort from pet and parental help. Children at high risk of animal harm were more likely to be classified as insecure (p = 0.002). Drawings indicated secure children tended to feel closer to mothers (p = 0.014) and siblings (p = 0.007), while pets’ proximity did not vary according to attachment strategy. Although insecure children scored lower on mentalising (p = 0.013), caregiving behaviour (p = 0.028) and parental help (p = 0.002), both groups similarly used pets as sources of comfort. There were no differences between attachment patterns on SAPS but there were differences for CAHB scores (p = 0.048). Thus, although insecure attachment was an important risk factor for harming animals, secure and insecure children had similar capacity for bonding with their pets. These results have implications both for how we treat cases of childhood animal harm and for how we understand the supportive role pets can play in children’s lives.