The Qinghai-Tibetan Plateau (TP), also named the Roof of the World, is known as the highest and largest plateau with an average altitude of about 4500 meters. Due to the unique orographical characteristics and thermal forcing mechanisms, the TP possesses unique plateau climate and, to a great extent, contributes to climate change over the whole world, especially East Asia. Intensive studies demonstrate that the warming pattern in TP exceeds that in other regions within the northern hemisphere and also the same latitudinal zone. In term of this result, the TP is regarded as one of the most sensitive regions to global climate change. However, the TP meteorological network is very sparse because of complex terrains and difficulties encountered in installing and maintaining the meteorological instruments. ERA-Interim has been a reanalysis of the global atmosphere covering the data-rich period since 1979, and is continuing in real time. ERA-Interim is the third-generation reanalysis product of European Centre for Medium-Range Weather Forecasts (ECMWF), and it uses an improved data assimilation system and an improved forecast model compared with ERA-40. Investigations have shown that the ERA-Interim could well capture the temperature patterns and it is very reliable for climate change research. In terms of their spatiotemporal distribution and variation characteristics, this paper analyzes the ERA-Interim precipitable water vapor (PWV) and surface temperature products at the spatial resolution of 0.125°*0.125° during 1979 and 2014. Moreover, the relationship between PWV and surface temperature changes was also investigated, using the empirical Orthogonal Function decomposition (EOF) method, correlation analysis and spectrum analysis. First, the EOF method is used to characterize the dominant spatial pattern and compact the representation of PWV and surface temperature. As a popular analysis tool in climate research, the EOF method is maximally efficient in retaining as much information of the data set as possible for as few degrees of freedom as possible through a linear combination of the original variables. Second, the correlation analysis is used to quantify the association between the PWV and surface temperature time series. A strong, or high, correlation means that two time series have a strong relationship with them and vice verse. Third, the spectrum analysis is the analysis of a spectrum of frequencies or related quantities. FFT (Fast Fourier Transform) method is adopted to analyze the cycle period of PWV and surface temperature in this paper. During the past 36 years, slight ascending pattern of PWV and significant ascending pattern of surface temperature were detected. According to previous researches, the increment of water vapor over plateau may originate from the evaporation increment of melting glaciers and snow. The warming pattern over TP in summer is the most significant and the warming pattern at higher altitudes is higher than that at the lower in winter. Evident seasonal variations appear in both PWV and surface temperature product of ERA-Interim since the transport of water vapor is associated with the monsoon. Moreover, the EOF analysis shows that time series of the first and the second patterns of PWV and surface temperature are significantly correlated. The correlation coefficient between the first PWV pattern and the fourth surface temperature pattern is up to 0.9. The correlation coefficient between these two patterns and the DEM (Digital Elevation Model) are 0.74 and 0.6 respectively. As the result the first PWV pattern can be regarded as the elevation component. Last but not least, an apparent 3-year cycle period is identified for both PWV and surface temperature series by FFT. And the consistent cycle period could justify the water vapor-temperature feedback.