Cognitive remediation is a promising pathway for ameliorating cognitive impairment of patients with schizophrenia. Here, we investigate predictors of improvement in problem-solving ability for two different types of cognitive remediation - specific problem-solving training and training of basic cognition. For this purpose we conducted a re-analysis of a randomized controlled trial comparing these two training approaches. The main outcome measure was improvement in problem-solving performance. Correlational analyses were used to assess the contribution of clinical, cognitive and training-related predictors. In the problem-solving training group, impaired pre-training planning ability was associated with stronger improvement. In contrast, in the basic cognition training group antipsychotic medication dose emerged as a negative predictor. These results demonstrate that predictors for successful cognitive remediation depend on the specific intervention. Furthermore, our results suggest that at least in the planning domain patients with impaired performance benefit particularly from a specific intervention.