Last years an optical thermometry became a mature science providing scientific basis for the remote temperature reading. Optical thermometry can be used in a wide range of applications, where the sensing requirements cannot be fulfilled by classical contact thermometers due to their limitations. A lot of effort has been made to develop efficient luminescence materials, read-out procedures, and temperature-dependent parameters. Here, single and multiparametric temperature sensing within 293–473 K were successfully demonstrated based on the monitoring of Pr3+-doped LaVO4 emission spectra. Thermometric performances of different sensing strategies, namely ratiometric technique, principal component analysis and partial least squares method, were compared in terms of thermal sensitivity and temperature uncertainty. Principal component analysis enhances thermal sensitivity by up to three times compared to the single parameter ratiometric technique at 293 K. The partial least squares method achieves a superior temperature uncertainty of 0.5 K. Among the samples studied, LaVO4:Pr3+ 1.0 at% is shown to be more suitable for optical thermometry than the 0.1 at% counterpart. The results obtained show the benefits of the advanced multivariate data analysis (search combination moving window interval partial least squares) for possible practical applications.