Abstract: With the rapid development of urbanization, the environments of weather stations have suffered to varying degrees. The representativeness and accuracy of observational data are affected at some urban stations because the surrounding environment outside the observation field is very complex, and the impact on meteorological factors cannot be ignored. Using these observed data may lead to partial or even wrong conclusions. To solve this problem, population-based methods are proposed to determine the type of stations, without fully considering the meso-scale environment. Therefore a new method is introduced to determine the type of stations by analyzing the land use within 20km.Previous studies usually focus on geographical area or all stations of country, but in order to investigate urbanization effects, it suggests choosing weather stations within the same climatic zone. The flat topographic and booming industry in Huang-Huai-Hai Plain provide favorable conditions for studying meso-scale impacts on temperature measurement. Using a temperature dataset of 199 stations during 2005—2007 and corresponding 20 km environmental survey data in 2007, the meso-scale environmental impacts on surface temperature are investigated. The data is revised based on latitude, longitude and altitude. The original observations without homogenization can reflect the impact of environment on the temperature because no station moved in this area during those yeares. This stations in the residential areas and farmland stations are compared by the average total annual value of the anomaly, the yearly value, the monthly value and the hourly value.The average observed temperature in residential area stations is 0.662℃ higher than that in farmland stations. The buildings around the weather stations have much effect on the minimum temperature measurements, while affect little on the maximum temperature measurement. The environmental impact in descending order is caused in winter, spring, autumn and summer. The temperature measurements at different times are affected by the surrounding buildings to varying degree. The impact on descending order is 02:00, 20:00, 08:00 and 14:00.In addition, the difference between two-type stations has also been analyzed when the residential and farmland environment within 20km range account for 60%, 70%, 80% of the whole area. It's found that with the typical underlying surface ratio grows, the temperature difference between the two-type stations becomes larger.