Due to climate change, extreme temperature events have become increasingly frequent in recent years. Adapting to extreme events is necessary to reduce their adverse effects on various sectors like agriculture, water resources, health, and ecosystems. For this, it is crucial to understand the extreme temperature characteristics. Temperature Duration Frequency (TDF) curves are useful for analyzing extreme heat events as they provide a relationship between the maximum temperature across various durations and their frequency. A non-stationary approach is used to account for external climate forcing, in which covariates (i.e., Global warming, humidity, and time) that cause extreme temperature changes are used for modelling. The non-stationary Generalized Extreme Value (GEV) analysis is conducted for Ahmedabad in India. Eight non-stationary models are created for each duration using three covariates and their combinations. For comparison, a stationary model is also developed. The analysis reveals a non-stationarity component in the AM temperature series of Ahmedabad city. Furthermore, it is noted that the stationary GEV model consistently underestimates the occurrence of extreme temperature events across all durations and return periods, as indicated by the comparison of temperature intensities.