A non-transferred plasma torch is a device used to generate a steady thermal plasma jet. Plasma torches have the potential to replace fossil fuel burners used as heat sources in the process industry. Today, however, the available plasma torches are of small scale compared to the power used in the burners in the process industry. In order to understand the effects of large scales on the plasma flow dynamics, it is essential to understand the operation of the plasma torch under different operating conditions and for different geometries. In this study, the analysis of a non-transferred plasma torch has been carried out using both computational and experimental methods. Computationally, the magnetohydrodynamic (MHD) equations are solved using a single-fluid model on a 2D axisymmetric torch geometry. The experiments are performed using emission spectroscopy to measure the plasma jet temperature at the outlet. This paper explains the changes in the arc formation, temperature, and velocity for different working gases and power inputs. Furthermore, the possibilities and disadvantages of the MHD approach, considering a local thermal equilibrium, are discussed. It was found that in general, the computational temperature obtained is supported by the experimental and equilibrium data. The computational temperatures agree by within 10% with the experimental ones at the center of the plasma torch. The paper concludes by explaining the significant impact of input properties like working gas and power input on the output properties like velocity and temperature of plasma jet. © 2023 Author(s).