Petroleum accumulations may coincide with either positive or negative temperature anomalies, which are conventionally detected using in situ temperature measurements made in shallow boreholes 1‐3 m deep. Data gathered in this way, however, can be sparse and costly, and may require intensive fieldwork over a long time period. This article explores the possibility of detecting thermal anomalies associated with petroleum entrapment using satellite‐derived land surface temperature data. For this aim, a robust correction scheme based on a physically‐based land surface model was applied to night‐time kinetic temperature data derived from NASA's ASTER instrument. The numerical model, known as SKinTES, attempts to simulate diurnal effects and to remove them from the measured temperature data to yield a residual temperature anomaly map. The performance of this methodology was tested over the Alborz oilfield located on an anticline of the same name in the Qom region of Central Iran. The study area has an arid to semi‐arid climate and the surface geology is dominated by outcrops of the Lower Miocene Upper Red Formation. The modelling approach used successfully highlighted several negative temperature anomalies over the oil‐bearing parts of the Alborz structure. In comparison to the uncorrected data, the anomalies were shown to be highly enhanced in both spatial and magnitude terms. In addition, time series analysis indicated that the temperature anomalies were consistent over time. The authenticity of the anomalies was confirmed by a suite of in situ temperature measurements made at shallow boreholes. In conclusion, a unifying framework is proposed to explain the occurrence of both negative and positive temperature anomalies over petroleum accumulations. The new modelling and correction scheme is expected to broaden the application of remote sensing land surface temperature data not only in petroleum exploration but also in other types of geothermic investigations including geothermal exploration.