Post-stroke depression (PSD) is the most common neuropsychiatric disorder after a stroke, with an incidence rate of 20% to 60%. PSD is not only associated with higher mortality rates, poorer recovery, more obvious cognitive impairments, greater economic burdens, and lower quality of life, but also brings additional medical expenses and care pressure to families. Society also needs to bear higher medical costs. Currently, the early diagnosis of PSD is difficult, which may lead to poor prognosis after stroke. This study aims to utilize machine learning technology to integrate multi-dimensional indicators of patients with ischemic stroke, establish a risk prediction model for PSD, and assist in early, accurate, and individualized assessment of PSD risk in clinical practice.