In recent years, the interest in the design, construction and utilization of rubber tracks for agriculture and earth moving machinery has increased considerably. The development of such types of tracks was initiated by the efforts to invent a more environmentally friendly vehicle-terrain system. These tracks are also the result of the continuous effort to develop more cost-effective traction systems. A rubber-surfaced and friction-based prototype track was developed and mounted on the patented modification of a new Allis Chalmers four wheel drive tractor. The track is propelled by smooth pneumatic tyres by means of rubber-rubber friction and the tractive effort of the track is mainly generated by soil-rubber friction between the rubber surface of the track elements and terrain. The experimental track layer tractor, based on an Allis Chalmers 8070 tractor (141 kW) was tested on concrete and on cultivated sandy loam soil at 7.8%; 13% and 21% soil water content. The contact pressure and the tangential force on an instrumented track element, as well as the total torque input to one track, was simultaneously recorded during the drawbar pull-slip tests. Soil characteristics for pressure-sinkage and friction-displacement were obtained from the field tests by using an instrumented linear shear and soil sinkage device. By applying the approach based on the classical bevameter technique, analytical methods were implemented for modelling the traction performance of the prototype track system. Different possible pressure distribution profiles under the tracks were considered and compared to the recorded data. Two possible traction models were proposed, one constant pressure model, for minimal inward track deflection and the other a flexible track model with inward deflection and a higher contact pressure at both the front free-wheeling and rear driving tyres. For both models, the traction force was mainly generated by rubber-soil friction and adhesion with limited influence by soil shear. For individual track elements, close agreement between the measured and predicted contact pressure and traction force was observed based on the flexible track model. The recorded and calculated values of the coefficient of traction based on the summation of the traction force for the series of track elements were comparable to the values predicted from modelling. However, the measured values of drawbar pull coefficient were considerably lower than the predicted values, largely caused by internal track friction in addition to energy dissipated by soil compaction. The tractive efficiency for soft surface was also unacceptably low, probably due to the high internal track friction and the low travel speeds applied for the tests. The research undertaken identified and confirmed a model to be used to predict contact pressure and tangential stresses for a single track element. It was capable of predicting the tractive performance for different possible contact pressure values.