Based on daily discoloration percentage and meteorological datasets, the meteorological conditions of cotinus coggygria leaves discoloration (CCLD) are studied and the staticstical prediction model is built. The results are as follows: (1) CCLD has a close relationship with temperature variation in the near several days, and the reliability of CCLD Day forecast by meteorological threshold is high. (2) The mean CCLD Day is between October 4 and 5, the earliest CCLD Day is September 27, and the latest one is October 13. (3) There is significant negative correlation between CCLD Day and the mean temperature (also minimum temperature and maximum temperature) in July. It should be noticed that the annual increment of CCLD Day has significant positive correlation with that of precipitation in July. (4) Choosing the annual increment of mean and maximum temperature in July as predictors, the statistical prediction model based on meteorology is built, and the mean error is around 3 days. (5) According to the 2011 independent sample verification and the 2012 forecast experiment, the prediction model of CCLD Day is proved to be usable.