Abstract: Protected horticulture is required for the high positioning accuracy and stability of mobile platforms. However, the electric drive mobile platforms can be often confined to object obstruction under the natural scenes. In this study, a high-precision positioning system was designed using an ultra-wide band (UWB) and inertial measurement unit (IMU). Additionally, an adaptive robust combination positioning algorithm (ARCPA) with Kalman filtering was developed to effectively improve positioning accuracy and stability. Firstly, a multi-sensor integrated positioning was constructed to combine UWB and IMU. Secondly, a differential distance-based sliding window was introduced to precisely detect nonline of sight (NLOS) errors in the positioning data of UWB. Thirdly, a cosine restart function was used to dynamically adjust the fading factor in real time, in order to reduce the interference of noise fluctuations in Kalman filtering. The accuracy and robustness of the filter were enhanced after optimization. Finally, a position dilution of precision (PDOP) weighted robust factor was incorporated to improve the resilience of the positioning system against noise and model disturbances under-protected horticulture scenes. An electric drive mobile platform and real-time positioning were constructed to verify the effectiveness of the integrated system. Then the simulation and real experiments of vehicle positioning were conducted in the typical protected horticulture. The experimental results demonstrate that the root-mean-square errors before and after UWB ranging correction under line-of-sight condition were 119.0 and 49.0 cm, respectively, and the ranging accuracy was improved by 58.8% after correction. The positioning error of the integrated positioning was only 6.63 cm, which was 81.92% higher than that of the single sensor (36.68 cm). The maximum error of positioning accuracy was also reduced by 88.89%. Therefore, this integrated positioning significantly reduced the impact of NLOS and geometric error on the positioning accuracy of the electric drive mobile platform for protected horticulture. Moreover, the root-mean-square error and the maximum positioning error of ARCPA were 17.70 and 44.44 cm, respectively. Furthermore, the accuracy of the improved system increased by 50.15%, 74.63%, 60.01%, 60.88%, and 28.34%, respectively, compared with the Kalman filtering, unscented Kalman filtering, extended Kalman filtering, particle filtering, and adaptive robust Kalman filtering. Correspondingly, the maximum positioning error was then reduced by 66.83%, 73.67%, 69.07%, 71.2%, and 50.50%, respectively. The integrated positioning significantly improved the robustness of the electric drive mobile platform for protected horticulture. The finding can also provide a specific theoretical basis and plan for the positioning and navigation in a protected horticulture environment.