An ultra wide band indoor positioning platform was developed to improve the positioning accuracy of agricultural vehicle in a greenhouse environment. Firstly, based on Ubisense radio frequency chip, a positioning platform tag and a base station were designed. The position data were obtained by location ranging method named time difference of arrival (TDOA). Then, in order to get accurate positioning information and meet the needs of precise positioning for agricultural vehicle in greenhouse, K-means clustering and segmentation processing method were applied for the optimization calculation of static and dynamic positioning data respectively. In this procedure, the measurement errors, which were caused by propagation characteristics of pulse signal and the single ranging method, were decreased. Finally, static and dynamic positioning tests were conducted under the greenhouse environment respectively. The experimental results showed that the optimized static positioning accuracy was 0.0634m. Meanwhile, the probability of the dynamic positioning accuracy, which was no more than 8cm gotten by virtue of the dynamic precision positioning algorithm can be increased by 3.7 times than that of the original Ubisense positioning platform. Hence, the platform built can meet the practical application requirements. The research provided a reference for the positioning, navigation and intelligent refinement operation of agricultural vehicle in the greenhouse.