The objective of this study is to determine the viscoplastic parameters and their uncertainty in ceramic materials at high temperatures by applying Bayesian inverse analysis to pyroplastic tests. Pyroplastic deformation is the bending of a ceramic specimen caused by gravity during heat treatment, and it is a measure of the loss of shape of a product during firing. Pyroplastic mechanical behaviour can be described by viscoplastic constitutive equations. Two different viscoplastic potentials have been proposed, each depending on either two or three parameters. The optimum values of the parameters and their corresponding uncertainty were determined using Bayesian inverse analysis. The study was carried out on industrial porcelain tiles, obtained from kaolinitic–illitic clays, quartz, and sodium feldspar. The experimental programme consisted of twelve bending tests performed at five different temperatures and three applied loads. The evolution of maximum displacement versus time was measured. The specimens were modelled using finite element analysis in COMSOL considering their viscoplastic behaviour. To simplify the material dependence by reducing the experimental data to one dimension, a non-dimensional physical subrogated model has been developed. The subrogate model has been used to determine the optimum viscoplastic parameters minimizing the differences between the experimental data and the numerical simulations. A Bayesian methodology has been applied to obtain the statistical posterior distribution of the parameters. The final output of the procedure is a joint probability distribution of the viscoplastic parameters.