The use of elevated temperature or temperature programming in liquid chromatography provides several advantages such as fast analysis, increased efficiency, a change of selectivity and an increase of the elution strength of the mobile phase. Method development in high- temperature liquid chromatography is usually governed by trial and error although a systematic approach is preferred. Therefore, it was investigated whether the empirical linear elution strength (LES) retention model can be adapted from temperature-programmed gas chromatography (GC) to temperature-programmed liquid chromatography (LC). It was found that by means of the LES model, retention times of selected steroids and polycyclic aromatic hydrocarbons can be precisely predicted depending on a simple linear temperature gradient in LC. An average relative error of less than 2% of predicted retention times was observed. Moreover, the influences of column chemistry, inner column diameter and composition of an isocratic mobile phase were studied. Because of these findings, the LES model was further extended in order to predict more complex segmented temperature gradients. For these gradients, the retention times of sulfonamides could be predicted precisely with an average relative error of 2.2%. The LES model in GC permits isothermal retention time predictions on the basis of temperature-gradient measurements. This approach was also employed in liquid chromatography and it is shown that this assumption cannot be transferred to temperature-programmed LC. Because of the need to predict isothermal retention times, predictions based on a plot of the natural logarithm of the retention factor were tested for temperature dependency. It was found that a plot of the natural logarithm of the retention factor versus temperature yields reliable isothermal retention time predictions. In order to improve the accuracy of retention time predictions based on temperature gradients even further, a second compound specific model parameter was also calculated temperature dependent. Using this approach, the relative error of retention time predictions of multi-step temperature gradients can be decreased to around 1.5%. Concurrently, a new experimental design was introduced which permits isothermal predictions on the basis of only four temperature-gradient input measurements. Moreover, a set of recommendations to assist the practitioner during method development in HT-HPLC was established. Finally, the linear solvent strength and the linear elution strength retention model were combined in order to predict simultaneous solvent and temperature gradients in LC. An average relative error of 0.6% of predicted retention times was observed. On the basis of the present work, temperature gradients can now be incorporated in systematic method development in liquid chromatography.