Abstract: The paper aims to improve the forecast level of fog, Based on the daily meteorological data obtained from Jinan National Surface Meteorological Observation Station during 1961-2014, we analyzed the variation trend of fog days used the linear trend estimation method and assess the abnormal conditions of fog days by using the anomaly evaluation criterion.The results showed that: there were decreasing trends for both annual and seasonal variation from 1961 to 2014 with a linear tendency rate of -1.362 d/10 a. The date of the annual average fog days was 16 days, the minimum value was 5 days appeared in 1969、2008 and 2010, and the maximum value was 39 days in 1964. There was 1 year which was significantly less than others. There were 5 years which were less than others. There were 41 years which were normal. There were 2 years which were more than others. There were 2 years which were significantly more than others. There were 3 years which were obviously more than others. For the seasonal variation linear tendency rate, the peak value was -0.444 d/10 a in fall, and the lowest value was -0.211 d/10 a in winter. It was 1.5 and 6.5 days for the mean seasonal fog days in spring and winter, respectively, which were the minimum and maximum date in the four seasons. The number of normal years was 50 for the fog days in spring, and spring was the season with the least number of abnormal years; while it was 34 for the number of normal years in winter, and winter was the season with the biggest number of abnormal years.