How well children remember negative events is not fully understood. Previous research has failed to simultaneously test memory and perceptions of memory for both negative and positive events. Children (n = 38, 7–17 years) recruited from a hospital following accidental injury were tested for their memory of an injury-producing accident (negative event) and a positive event (unexpectedly receiving a $50 gift voucher). Objective accuracy of memory, memory quality characteristics (e.g., how coherently the event was recalled), children's judgments of their memory (meta-cognitive), and posttraumatic stress (PTS) symptoms were assessed 2 months post-injury. Children's memories for their experiences were verified using witness/parent reports. Memory quality characteristics of children's free recollection were coded with a previously used standardized system. Overall, children showed high levels of accuracy for both events, with little degradation over time. High PTS children showed little evidence of deficits in coherence or organization in their narratives relative to low PTS children. Although in some instances high PTS children judged their memory quality to be poor compared to low PTS children, this depended on how this was assessed (e.g., self-report questionnaire vs. coded narratives). In terms of limitations, it is unclear whether the findings will generalize for memories of repeated events. Witness verification of the accident details itself could be prone to error. In conclusion, the findings are broadly supportive of the proposal made by theorists who argue that trauma memories are recalled no less accurately than other distinctive memories. The role of meta-cognitive elements of children's memory and reporting in PTS is less clear.