The contact resistance of the pantograph system directly reflects the stability and reliability of the system operation. In order to solve the problem that the conventional calculation accuracy of contact resistance cannot meet the engineering requirements, this paper proposes a contact resistance prediction model based on the joint construction of variational mode decomposition and extreme learning based on the improved whale optimization algorithm, which has the advantages of strong global search ability, fast convergence speed and high accuracy. Firstly, the improved whale algorithm is used to optimize the K value, penalty coefficient and the implicit layer weight and threshold of the extreme learning machine respectively, and the experimental data of the accessible resistance of the optimized variational mode decomposition are optimally decomposed to obtain the eigenmode function with better characteristics. Secondly, the optimized extreme learning machine effectively avoids the blindness of the support vector machine in selecting parameters. Finally, a prediction model is established to predict each modal component, and the prediction results of each modal component are obtained and reconstructed to obtain the final prediction results. Experimental results show that a new prediction model proposed in this paper has very good applicability to the prediction of nonstationary timing of contact resistance.