This paper presents a comprehensive structural orientation model to facilitate a hypothesis for the percolation threshold of nanotube-polyurethane systems, utilising Monte Carlo simulations. It considers a representative volume for nanotubes of diverse sizes and aspect ratios, uniformly dispersed at random positions and orientations. Continuous conductive networks formulated with the representative elements were identified for various filler contents, modelled and predicted percolation threshold of 0.19%. Experimental percolation threshold obtained was 0.21%, which determines the reliability of the model and furthermore, the model is broad and can, therefore, be extended to any nanocomposite, to predict the percolation threshold. The shape memory effect of the nanotube-polyurethane nanocomposite was evaluated for thermal and electrical stimuli, and the recovery efficiencies arrived to 95 and 98% respectively.