In first-episode schizophrenia the advantage of new atypical neuroleptics compared to low-dose haloperidol as well as the indicated duration of neuroleptic maintenance treatment has still to be based on empirical evidence. Accordingly, a multi-center study on the optimization of acute and long-term treatment in first-episode schizophrenia is currently being carried out as part of the German Research Network on Schizophrenia. This paper reports on the design, methods and preliminary results of the two-year randomized double-blind study comparing risperidone and low-dose haloperidol within the framework of psychological interventions. In the second treatment year, relapse rates under continued neuroleptic treatment are compared with those under stepwise drug withdrawal substituting instead prodrome-based early intervention (intermittent treatment). As to the results, by November 2003 142 first episode patients (ICD-10 F20) have been included in the long-term study. One-year relapse rates were very low (3.8 %). On average, symptoms as well as drug side-effects decreased steadily under maintenance treatment. Although compliance on average was high, about 60% of the patients dropped out during the first study year. More pronounced psychopathology, (neurological) side-effects, lower compliance at study entry and absence of psychological treatment seemed to enhance the risk for drop-out. In conclusion, treatment in first episode schizophrenia is effective under both (further on blinded) neuroleptics; however these patients are at high risk for treatment drop-out. This emphasizes the need for a special support program.