This systematic review aimed to examine whether neurobiological methods, or other methods independent of clinical judgment, have been investigated to assist decision making in forensic mental health services and, if so, whether this may be a useful strategy for predicting outcomes. OVID-Medline, Embase, and PsychInfo (inception–January 2015) were searched, limiting to English and human studies, using terms relating to “predict,” “outcome,” “psychiatry,” and “forensic” to identify primary research articles reporting on predictors of outcome in forensic mental health services not reliant on clinical judgment/self-report. Fifty studies investigating demographic, neuropsychological/neurophysiological, and biological predictors were identified, reporting on 3 broad outcomes: (i) inpatient violence, (ii) length of stay, (iii) reoffending. Factors associated positively, negatively, and showing no relationship with each outcome were extracted and compiled across studies. Of various demographic predictors examined, the most consistent associations were between previous psychiatric admissions and inpatient violence; a more “severe” offense and a longer length of stay; and young age and reoffending. Poor performance on tests of cognitive control and social cognition predicted inpatient violence while a neurophysiological measure of impulsivity showed utility predicting reoffending. Serum cholesterol and creatine kinase emerged as biological factors with potential to predict future inpatient violence. Research in this field is in its infancy, but investigations conducted to date indicate that using objective markers is a promising strategy to predict clinically significant outcomes.