Metacognitive theory proposes that depression is caused by excessive rumination, which is in turn maintained by maladaptive positive and negative beliefs about rumination (“metacognitions”) and reduced executive control. Moreover, the metacognitive model asserts that metacognitions are maintained by prolonged depression symptoms. However, no studies have tested the metacognitive model of depression prospectively in a clinical population. Currently remitted adults with recurrent depressive disorder (N = 105) reported depression symptoms at five time points over a 12-month period. Based on this data, we used latent growth modelling to estimate depression levels and symptom trajectories. Positive metacognitions were associated with rumination, while negative metacognitions and rumination predicted higher depression levels, but not symptom recurrence. Moreover, depression levels and symptom recurrence predicted positive and negative metacognitions, as well as rumination. There was no association between metacognitions and reduced executive control. The present study lends partial support for the metacognitive model, but raises questions of the relevance of metacognitions as a proximal vulnerability marker for symptom recurrence.