Based on the concept of entropy in information theory, the entropy of meteorological elements is determined and used to analyze the uncertainty of the global temperature field anomaly from the climate state (1971—2000) on different time and spatial scales. It is found that the temperature climate entropy (CE) possesses a zonal distribution, increases from tropics to mid-high latitudes and has an obvious difference between the ocean region and the continent, thereby being able to distinguish the climatic zones properly. The temperature CE in low-mid troposphere increases with altitude increasing, while in extratropical the situation retains above 300 hPa but below 300 hPa the situation is reversed, and this feature is more obvious in high latituderegions. On the whole, the temperature monthly entropy (ME) is obviously dependent on season change: it is smallest in summer and largest in winter. Besides, there exists a distinguishable interdecadal period. Different monthly ME values from low atmosphere to high atmosphere each have an obvious five -to-ten year quasi-period oscillation. All the spatiotemporal characteristics and their relationships with annual temperature range verify the usefulness of the entropy in meteorology, and it is an effective method to measure the uncertainty of the meteorological elements.