ABSTRACTIn this paper, we investigated changes in heat and cold waves in Romania over the period 1961–2015 by employing a new and superior approach. It consists in using excess heat factor to identify heat waves and excess cold factor to identify cold waves. Five indices were calculated and then analysed for both heat waves and cold waves resulting in a set of ten indices. Indices for heat waves were analysed for the extended summer season (May–September), whereas those for cold waves were assessed for the extended winter (November–March). The intensity threshold was set to be equal or above the 90th percentile for heat waves, and equal or below the 10th percentile for cold waves, while the duration threshold for both heat and cold waves was of at least three consecutive days. For a better comparison with other studies conducted worldwide, and to get more information from the data sets, the percentile thresholds for heat and cold waves identification were calculated based on three reference periods: 1961–1990, 1971–2000, and 1981–2010. Trends were calculated using ordinary least square method, whereas statistical significance was assessed by the t‐test. The main results indicated that changes are more substantial in the case of indices calculated based on excess heat factor compared to those based on excess cold factor, suggesting that the warming process is more reflected in heat waves rather than in cold waves. Thus, heat waves became more frequent, longer, and more intense, while cold waves became less frequent, but more intense. When the reference period for percentile threshold calculation was changed from the earliest to the most recent ones, the frequency of increasing and significant increasing trends decreased for some of the heat wave indices, while for the cold wave indices the significant downward trends increased.