Based on the potato yield data in Hebei Province from 1983 to 2019, the average temperature, precipitation and sunshine hours every ten days from 21 meteorological stations during the same period, the meteorological key factors influencing potato yield were determined by means of factor puffing and correlation analysis. The forecast model of potato yield was established and this model was tested during 1983-2015 and was applied during 2016-2019. The results show that potato yield in Hebei Province was sensitive to temperature and sunshine from July to September, and the effect of temperature on potato yield was greater than sunshine hours. The average trend backtesting accuracy ranged from 75.7% to 91.7% for different initial forecast time, the average backtesting accuracy of forecast potato yield ranged from 83.6% to 91.4%, and the forecast accuracy of potato yield during 2016-2019 ranged from 83.5% to 95.3%, and with the backward of initial forecast time, the forecast accuracy improved gradually.