Abstract: In order to improve the quality of hourly observations with long-term series at regional automatic observation station, as well as enhance its application value in coping with and adapting to climate change, the homogenous hourly temperature data of regional automatic weather station over 2008-2023 has been developed, taking Tianjin area as an example. This dataset eliminates erroneous values from the observational time series objectively, it has also effectively mitigated and evaluated the systematic bias impact caused by non-climatic factors such as station relocation, upgrade of observational instrument elements, and collector renewal on the time series of regional automatic observation station. Firstly, in the data quality control, 148 errors of 15 stations detected by internal consistency for the three temperatures were set to the missing value or corrected by manual verification. And 19 errors of 3 stations, 21 errors of 18 stations and 66 errors of 31 stations were set to missing values detected by climate outlier, for hourly mean temperature, maximum and minimum temperature, respectively. Correspondingly, for the detecting results from the spatial consistency, 5 errors of 3 stations, 2 errors of 2 stations and 1 error of 1 station were also set to missing values. Secondly, in the homogenization analysis, 6 stations were eliminated by means of the maximum penalty F test (PMFT) combined with station metadata, which had statistically significant breakpoints supported by accurate metadata. Finally, the developed hourly temperature data of regional automatic observation station was verified relatively reliable by comparing with the corresponding observations of the national stations at 11 administrative regions in Tianjin. In addition, using the hourly data of 106 regional automatic observation stations, it is shown that the increasing trend and the amplitude of the maximum temperature in Tianjin during the recent 10 years is relatively the most significant, followed by the mean temperature, and the minimum temperature is relatively the least, especially in autumn, 100% of its regional automatic observation stations show an increasing trend in the maximum temperature, among which 63.2% are statistically significant (?? = 0.05), with the amplitude range of 1.316-3.760°C/10yr.