We used PDA devices and an experience sampling technique to assess participants’ negative mood and thoughts as they engaged in their normal daily routines over the course of a week. We then calculated each person’s own unique relationship between mood and thoughts, and used this index of cognitive reactivity to predict depressive symptoms at 6-month follow-up. Participants who demonstrated a stronger link between their momentary negative mood and negative cognitions reported more depressive symptoms at follow-up than those who had a weaker relationship between mood and cognitions. Further, this cognitive reactivity index was a better predictor of follow-up depressive symptom scores than initial depressive symptoms, dysfunctional attitudes, average experienced negative mood or thoughts, or variability of negative mood or thoughts. These results are consistent with earlier findings and build on previous research by demonstrating that naturally occurring cognitive reactivity is predictive of future mood disruptions.