•Comprehensive multicomponent geothermometer with integrated optimisation process for aluminium concentration, pH value, as well as boiling and dilution.•Development of a universally valid mineral set for worldwide applicability of mulT_predict.•Development of an outlier removal procedure for more precise temperature estimation.•Benchmarking the optimisation process of mulT_predict by perturbation of a synthetic brine. In this study, we introduce MulT_predict as a fully integrated solute multicomponent geothermometer, combining numerical optimisation processes for sensitive parameters to back-calculate to chemical reservoir conditions. This results in a state of the art geothermometer, providing an accurate reservoir temperature estimation validated by geothermal borehole measurements on a worldwide scale. In addition, a universally valid mineral assemblage for an unknown reservoir composition is developed, focusing on worldwide applicability. Using the evolved methodology, the limits of the optimisation processes are determined by using a synthetic brine (150 °C, pH 6, aluminium concentration 0.003 mmol/l) and successively perturbing its geochemical equilibrium state. Individual back-calculation of reservoir conditions lead to valid temperature estimations of 145 °C, 3.4% lower than the initial temperature while a simultaneous and interdependent optimisation reconstructs the sensitive parameters even more precisely with a deviation of 0.056 for the initial pH value, and 0.164 µmol/l for the aluminium concentration.