There has been a move in the kiwifruit industry to pollinate by machine, allowing orchardists flexibility in timing, and other advantages. The traditional pollinators, bees, are unreliable and are susceptible to diseases such as the varroa mite. No measurements have been done on the efficiency of single-flower collection of pollen and the influence of airjet characteristics, to put this activity on a firm quantitative basis. For predictive purposes, a robust model of pollen behaviour around a single Green kiwifruit flower (Actinidia deliciosa) was built in this study by using a commercial computational fluid dynamics (CFD) package. The theoretical study first looked at the wind pollination of flower buds at different opening stages. Within the range of draft velocity recorded in the orchards, stigma are predicted to capture more pollen in a wind approaching from the front than from the side and back. Also, the pollen deposition on stigma increases as the bud opens wider and loses its petals. An estimation based on the predicted pollen collection efficiency indicated that wind pollination alone is insufficient to produce a minimum exportable fruit, which is in agreement with the observations reported in the literature. The question whether kiwifruit is a wind-pollinated plant is also discussed. Experimental checks of the modelling were carried out. Visualization of flow paths around a real kiwifruit flower in a wind tunnel compared well with the CFD predictions. The machine pollination employs a jet of air. CFD simulations of pollen-loaded air jets were carried out to study the effect of different jet-nozzle configurations, namely the jet direction onto the flower, nozzle-to-flower distance, diameter of nozzle and initial jet velocity. A frontal spray gives the most efficient pollen delivery to the stigma. Further gain is accomplished by introducing the pollen cloud closer to the flower, from a smaller nozzle and at higher initial jet velocity. These trends, except for that of nozzle size, have been validated experimentally through the pollen spraying of single real kiwifruit flower in the laboratory. Recommended ways of airjet spraying are suggested based on the CFD results. From the stigma collision efficiency, it is more advantageous to apply pollen when most of the flowers have lost their petals. Additional inclusion of an electrostatic influence into the pollen-loaded air jet simulations indicate that corona charging, but not tribo-charging, gives enhanced pollen collection by the stigma. The enhancement depends on the strength of electric field established between the nozzle and the flower. The spraying tests with tribo-charged pollen in laboratory indicated that an improved pollen collection may be possible with a greater charging of pollen. The benefit however was not realized consistently in all experimental conditions. The simulations assumed complete stickiness i.e. complete capture of pollen by the stigma on collision. This study also measured the force necessary to remove pollen in contact with stigma. The stickiness measurements of the stigma of Gold and Green kiwifruit flowers found that both flowers have equally sticky stigma. Stickiness increased with the aging of flowers. A reason is proposed related to the sugar in the exudate film. Overall, the variation of stickiness was statistically insignificant during the day. However, some older stigma were found to be stickier in the afternoon, which is in contrast to the general view that stigma are stickier in the morning. The gain in stickiness by the older stigma supports the recommendations for carrying out machine pollination after the petal fall. The last section of the work was done in an orchard, and studied the characteristics of large jet spraying (without pollen) at the T-bar and pergola trained vines. Recommendations on the designs of jet have been provided.