This systematic review aimed to determine whether the risk of extrapyramidal side effects (EPS) differed between antipsychotic drugs used in first episode psychosis (FEP). We identified 11 RCTs comparing two or more antipsychotics in FEP and reporting on EPS. All trials assessed one or more second generation antipsychotics (SGAs), one assessed chlorpromazine, one zuclopenthixol and seven trials assessed haloperidol. Assessment and reporting of EPS varied. Compared with one or more SGA comparators, haloperidol was associated with significantly higher rates/severity of parkinsonism (seven trials) and akathisia (six trials) and greater use of anticholinergics (five trials) and beta-blockers (two trials). Two trials with low-dose haloperidol (≤ 4 mg) showed significantly worse EPS outcomes versus a SGA. Two of four long-term haloperidol trials (≥ 1 year) found a higher dyskinesia-risk with haloperidol versus olanzapine and risperidone respectively; the remaining two trials found no difference (various SGA comparators). There was an EPS advantage for clozapine versus chlorpromazine (one trial) and risperidone versus zuclopenthixol (one trial). There was little evidence of EPS-differences between SGAs, possibly reflecting use of low doses. We conclude that SGAs offer an EPS advantage over FGAs in FEP though the evidence largely relates to comparisons with haloperidol. Standardized assessment and reporting of EPS would assist future research.