Stroke, a leading cause of global mortality and disability, often results in post-stroke cognitive impairment, which significantly affects survivors and increases the risk of dementia. Loneliness and reduced autonomous participation are common among post-stroke patients, yet their relationship with mild cognitive impairment remains underexplored. This cross-sectional study of 285 stroke patients assessed cognitive function, autonomous participation, and loneliness, using mediation analysis to examine their interrelationships. The findings revealed that risk factors for mild cognitive impairment included older age, chronic diseases, longer disease duration, memory decline, and loneliness, while autonomous participation was found to be protective. Significant correlations were observed: higher autonomous participation correlated with lower mild cognitive impairment (r = -0.384, P < 0.01) and lower loneliness (r = -0.394, P < 0.01), while loneliness was positively correlated with mild cognitive impairment (r = 0.431, P < 0.01). Mediation analysis identified loneliness as a partial mediator between autonomous participation and mild cognitive impairment, suggesting that addressing loneliness may improve cognitive outcomes in post-stroke patients. The study highlights loneliness’s mediating role in the autonomous participation-mild cognitive impairment relationship, suggesting clinical interventions targeting loneliness, particularly in low-autonomy patients, may mitigate cognitive decline. Limitations include single-center sampling and cross-sectional design, necessitating future longitudinal studies. These findings underscore the need for integrated care strategies addressing psychosocial and functional factors to improve post-stroke outcomes.