The paper presents a novel approach for controlling highly nonlinear systems using the combination of LPV framework and ultra-local model-based solution. Firstly, a new formulation of the ultra-local model is presented, by which the implementation-related issues can be overcome. Secondly, an extended state-space representation is introduced, which includes the nominal model of the considered nonlinear system and the effect of the ultra-local model. This extended state-space representation serves as the basis of LPV-based control design. In this way, the stability of the closed-loop system can be guaranteed, while the variation of the tuning parameter (α) of the ultra-local model can also be handled. The effectiveness and the operation of the proposed control strategy are demonstrated through a vehicle-oriented control problem.