Due to toughened CO2 emission standards the automotive industry is approaching a transition to zero-emission mobility. An accurate estimation of the stator temperature is an essential part of the motor controls and to protect the electric motor from overheating. An inaccurate estimation leads to bigger margins between estimated value and actual value to protect the electric motor which affects the performance. A literature study was performed which presented different methods that has been used in previous work to estimate parameters of an electric machine. The purpose of this master thesis is to find a suitable method and develop a temperature observer to estimate the stator winding temperature based on a resistance observer.In the development work three different methods were tested on a single-phase electric machine for the sake of simplicity. First method assuming steady state condition, second was the Forward Euler and third zero-order hold. The results showed that the zero-order hold method using a recursive least-square (RLS) optimization was performing the best in the higher frequencies but in the lower range the steady state RLS method was better which was also the method that was used on a three-phase electric motor model from a Nissan Leaf.The sensitivity analysis showed that the estimate of the stator resistance is sensitive to errors on the input variables and that the errors are highest in the mid-torque range and in the lower part of the speed range. The error on the input variables was arbitrarily chosen of +3 % and was added separately on the inductances and voltages of the d- and q-axis 𝐿𝑑, 𝐿𝑞, 𝑣𝑑, 𝑣𝑞, the rotor speed 𝜔 and the flux-linkage 𝜑. The variable that was affecting the resistance estimation the most was when a fault of 3% was introduced on the rotor speed 𝜔 which gave an error of more than -1200% on the estimated resistance. In nominal conditions, i.e. room temperature and constant stator temperature, the results showed an accurate resistance estimation of -0.4 – 1.4 % followed by an accurate temperature estimation