For many chronic stroke survivors, persisting cognitive dysfunction leads to significantly reduced quality of life. Translation of promising therapeutic strategies aimed at improving cognitive function is hampered by existing, disparate cognitive assessments in animals and humans. In this study, we assessed post-stroke cognitive function using a comparable touchscreen-based paired-associate learning task in a cross-sectional population of chronic stroke survivors (≥ 5 months post-stroke, n = 70), age-matched controls (n = 70), and in mice generated from a C57BL/6 mouse photothrombotic stroke model (at six months post-stroke). Cognitive performance of stroke survivors was analysed using linear regression adjusting for age, gender, diabetes, systolic blood pressure and waist circumference. Stroke survivors made significantly fewer correct choices across all tasks compared with controls. Similar cognitive impairment was observed in the mice post-stroke with fewer correct choices compared to shams. These results highlight the feasibility and potential value of analogous modelling of clinically meaningful cognitive impairments in chronic stroke survivors and in mice in chronic phase after stroke. Implementation of validated, parallel cross-species test platforms for cognitive assessment offer the potential of delivering a more useful framework for evaluating therapies aimed at improving long-term cognitive function post-stroke.