Objective To develop a nomogram prediction model for early screening of post-stroke depression （PSD）, and to provide an instrument for screening high risk patients. Methods Ten risk factors were collected from 259 stroke patients （the derivative group） within 7-14 days after stroke onset, and depression was measured using the Hamilton Depression Inventory and the Montgomery Depression Inventory in 8-10 weeks after stroke. The risk factors were determined using Chi-square test, Lasso regression and logistic regression to establish a nomogram prediction model, and the model was validated internally and externally (the validation group, 82 patients). Results The detection rate of PSD was 39.38% in the derivative group.Gender, marital status, number of coexisting diseases, lesion location, degree of neurological impairment, and ability to perform activities of daily living were independent risk factors for PSD (all P＜0.05).The nomogram prediction model based on the above-mentioned 6 independent risk factors had good discrimination (AUC value: 0.883 in internal validation and 0.849 in external validation) and accuracy (Hosmer-Lemeshow test:χ2=7.939,P=0.439 in internal validation, and χ2=3.538,P=0.896 in external validation). Decision curve analysis showed that the prediction model curve had clinical benefit when the threshold probability value was greater than 10%. Conclusion There is a high incidence of PSD in stroke patients 8 to 10 weeks after the onset of stroke.The nomogram prediction model can effectively predict the risk of early PSD and facilitate clinical delivery of targeted interventions.