Solute geothermometry often leads to a broad range and often inconsistent calculated reservoir temperatures, in particular when exploring geothermal systems, where only limited information (geology, borehole data etc.) is available. The application of different Na-K and SiO2 geothermometer, the most widely used methods, not uncommonly lead to deviations of results by up to 200 K for one sample. In this study, the most effective interfering factors for these geothermometer applications are identified. A multi-step approach is proposed, combining experimental and numerical methods with specific fluid characterization to quantify these factors and to transfer these findings to the natural system enabling the correction of temperatures to realistic in-situ values. Taking into account dilution with surface water, a chlorofluorocarbon concentration based mixing model was set up to correct analysed SiO2 concentrations to original in-situ concentrations. A numerical model was used to determine the in-situ pH, which is highly sensitive to silica solubility. Results from long-term laboratory equilibration experiments were evaluated to identify the reservoir type dependent equilibrated SiO2 polymorph. In the case of the Na-K geothermometer, it is shown that the Na+/K+ concentration ratio in fluids is obviously not unequivocally controlled by temperature but is also dependent upon reservoir rock composition. Thus, different reservoir lithologies lead to different equilibration states in terms of Na+/K+. This is obviously one reason for the existence of the large number of different Na-K geothermometers. By modelling the stability of the Na+/K+ ratio governing feldspars, albite and orthoclase, we suggest a method that reveals the Na+/K+ equilibration state for each fluid supporting the allocation of the appropriate geothermometer equation. The improvement procedure is demonstrated in a case study evaluating fluid data of geothermal springs from the Villarrica geothermal system, Southern Chile. It is shown that initially highly scattered results strongly converge after corrections, leading to a substantial improvement in in-situ temperature estimations with small deviations of <= 10 K between SiO2 and Na-K geothermometers. Also absolute temperature calculated for each spring in the study area, ranging from 84 to 184 degrees C agree well (within Delta T < 20 K) with results of multi component geothermometry temperatures reported in a previous work.