In response to environmental concerns, the composites industry is showing a growing interest in natural fibre biocomposites as an alternative to wood plastic composites and glass fibre thermoplastics. Albeit many years of research, the potential of these new materials has not being reached and the properties obtained are too often lower than expected. A main reason for this is because the natural fibre properties are variable and poorly characterised, and inefficient traceability makes it difficult to grade the fibres. When it comes to biocomposite manufacturing, short plant fibre composites have attracted the interest of the thermoplastic compounding industry but only a few companies have mastered the compounding step. Too often, the extruder is treated as a "black box" and the critical processing parameters have not yet been identified. As a result, the full potential of extrusion process for biocomposites is not exhausted. In addition, the critical parameters to determine the BET surface area values with IGC were identified and a protocol applicable to natural fibres was proposed.The series of extrusion trials brought new insights into the feasibility of large scale biocomposite extrusion. Statistical analysis showed a significant interdependence between all factors and particularly between the screw speed and the screw design. At both laboratory scale and i medium scale, fibre content was the dominant factor for the tensile strength and elastic modulus whilst screw speed and screw design affected to a lower degree the tensile properties. The fibre surface properties and fibre length distribution were also determinant for the biocomposite properties. For instance, the alkaline treated fibre reinforced polypropylene composites underperformed compared to the water washed fibre polypropylene composites although the former had a surface more energetically homogeneous and less polar. It is assumed that the higher fibre aspect ratio of the water washed fibres and their homogeneous fibre length distribution largely contributed to increase the composite performance.Finally, extrusion at industrial scale has been successfully performed and represents a major achievement of the thesis. However, significant amount of porosity was noticed in the extruded samples throughout the trials and further work is required to overcome this issue. Whilst the porosity detected in the samples questions the industrial-usefulness of some of the results, the contribution of this thesis to the development of natural fibre compounding capability at The University of Queensland and collaborating local industries was immense. The methodology and the lessons learned will undoubtedly be used to further optimise the extrusion process and produce better biocomposite materials.ii