The prodromal phase of Parkinson’s disease (PD) is characterised by many non-motor symptoms, and these have recently been posited to be predictive of later diagnosis. Genetic rodent models can develop non-motor phenotypes, providing tools to identify mechanisms underlying the early development of PD. However, it is not yet clear how reproducible non-motor phenotypes are amongst genetic PD rodent models, whether phenotypes are age-dependent, and the translatability of these phenotypes has yet to be explored. A systematic literature search was conducted on studies using genetic PD rodent models to investigate non-motor phenotypes; cognition, anxiety/depressive-like behaviour, gastrointestinal (GI) function, olfaction, circadian rhythm, cardiovascular and urinary function. In total, 51 genetic models of PD across 150 studies were identified. We found outcomes of most phenotypes were inconclusive due to inadequate studies, assessment at different ages, or variation in experimental and environmental factors. GI dysfunction was the most reproducible phenotype across all genetic rodent models. The mouse model harbouring mutant A53T, and the wild-type hα-syn overexpression (OE) model recapitulated the majority of phenotypes, albeit did not reliably produce concurrent motor deficits and nigral cell loss. Furthermore, animal models displayed different phenotypic profiles, reflecting the distinct genetic risk factors and heterogeneity of disease mechanisms. Currently, the inconsistent phenotypes within rodent models pose a challenge in the translatability and usefulness for further biomechanistic investigations. This review highlights opportunities to improve phenotype reproducibility with an emphasis on phenotypic assay choice and robust experimental design.