Introduction: The amount and state of vegetation in urban areas, especially in metropolises, play a crucial role in the urban ecosystem functions and life quality of its residents.Therefore, it is important to investigate its quantitative and qualitative changes at a local scale, especially within and around  big cities. The aim of this research is to investigate the temporal changes of the total vegetation cover using Normalized Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) indices of Landsat images over the Kermanshah metropolis and its eight districts during 1377-1400 period and its relationship with the weather parameters.
Material and Methods: Three to four Landsat images acquired during the spring season of each study year were downloaded for each year. After checking the geometric accuracy, the atmospheric correction was applied and vegetation indices, i.e. NDVI and EVI were generated. Using the city boundary map, the mean values of the mentioned vegetation indices within the city boundary were calculated as the total vegetation cover. From the images of each year, the image corresponding to the peak greenness was selected based on the maximum vegetation index values, and a time series of 24 peak greenness images was generated for each vegetation index. Time series analyses were applied using a parametric method i.e. Ordinary Least Squares (OLS) regression and non-parametric, i.e. Thiel-Sen and Mann-Kendall approaches. These analyses were repeated in each district, also, the weather data (i.e. precipitation and temperature) were used to explore their relationship with the changes in vegetation indices.
 Results: In most years, the Ordibehesht (i.e. mid-April to mid-May) images showed the highest vegetation cover values. Based on the NDVI, it can be stated that the overall vegetation cover in Kermanshah metropolis has experienced a weak decreasing trend  with slopes of -0.0013 and -0.001 obtained using the parametric Ordinary Least Squares (OLS) regression and the non-parametric Theil–Sen method, The results of the EVI index was similar. Based on the Mann–Kendall test (α = 0.05), with Mann–Kendall statistics of -1.27 and -0.52, the trends were not significant. According to the NDVI analysis, District 4, exhibited a non-significant increasing trend, while the other regions showed non-significant decreasing trends; only District 6 had a significant decreasing trend. The EVI analysis produced results similar to NDVI, with the difference that, in addition to District 4, District 5 also showed an increasing vegetation cover trend, and none of the decreasing trends were significant. The analysis of the relationship between changes in NDVI and EVI and weather variables with mean temperature and total precipitation one month and one year before the date of peak vegetation showed a positive correlation between vegetation cover changes and total precipitation, and a negative correlation with mean temperature. Weather data from one month before showed a stronger correlation with vegetation cover changes than data from one year before.
Conclusion: According to the results, the highest vegetation fraction and greenness is observed in Ordibehesht (i.e. mid-April to mid-May). A negative trend in vegetation greenness was observed, although, it was not statistically significant, that can be a warning alarm for city managers to implement a comprehensive plan to develop and preserve the vegetation fractions. District 4, also District 5 showed an increase in vegetation greenness. Historical Landsat satellite images are valuable data for investigating the qualitative changes of vegetation over long-term periods. Weather parameters are not always the main drives of vegetation changes, especially in urban areas that are strongly influenced by human activities, i.e. irrigation green space and construction activities. There is a space to explain mechanisms to investigate the role of these factors in the continuation of such studies.