This paper deals with the analysis of time trends for European average temperature anomalies for the period from 1655 to 2017 using a dataset based on the Central England Temperatures (CET) and other meteorological stations. The reason for this study is to determine first if long memory is present in the data, and then to estimate the time trends in a more rigorous way. The results indicate that the series display long memory behavior. Based on this and using recursive methods, we show that the time trend becomes statistically significant when the data from 2010 onwards is incorporated, indicating the importance of the last few years in the evidence for warming. It should be noted, however, that this last result may be a consequence of the statistical method employed and should, therefore, be treated with caution.