Abstract: In recent years, due to the transfer of rural labor and the increase of labor cost, there have be more and more researches and applications on the orchard platform which was used for assisting artificial works. The workspace is an important index to measure the organic integration of orchard platform and orchard planting mode. The researcher about kinematics and workspace analysis of agricultural machinery or agricultural robots was few. The values of the samples with the traditional Monte Carlo method are evenly distributed. In the course of the mapping process, the non-linear equation 'stretches' or 'compresses' the uniformly distributed intervals, which results in uneven distribution of the boundary position of platform reference point. The sparse point cloud and large 'noise' makes the workspace boundary inaccurate and thereby affects the optimization of mechanical parameters of orchard platform and other follow-up researches. In order to improve the utilization rate of orchard platform's workspace and make the operators more comfort, a kinematic model of orchard platform based on the D-H method was established to deduce the positive solution of the platform reference point kinematics of orchard platform and analyze the workspace. In view of the problems, such as workspace boundary fuzziness of orchard platform and point waste etc, an improved Monte Carlo method in which the domain of definition was of the U-shaped arcsine distribution was proposed, and a gridding algorithm was used to analyze the workspace generated by the ones before and after improved, then a cuboid was constructed which was divided into several subspaces along the direction of axis Z-Y-X contained all simulation points. The boundary definition therein was obviously improved, the number of boundary distribution points of workspace was increased by 23.74 %, and the area of the bottom and the top boundary layer was increased by 180.70 % and 102.69 % respectively. By taking the mean value of the workspace generated multiple times as the reachable workspace of orchard platform, the spatial difference coefficients in three directions were reduced 6.92 %, 8.99 % and 2.64 % respectively compareed with those before improved, which indicated that the workspace after improved was closer to the ideal workspace. Through the platform reference point position tracking test, the measured value of platform reference point basically agreed with the calculated value. When Z was 1 314.4 mm, the maximum difference of Y direction was 1.2 mm; when Y was 763.6 mm, the maximum difference of Z direction was 6.2 mm. The results showed that the kinematic model and the solution method were correct. On the actual workspace test of platform reference point, there was larger difference between the actual workspace and the ideal workspace of platform reference point, which was maximum in the direction Y and it was 10.27 %, but which between actual workspace and reachable workspace was less and they were respectively 0.27 %, 0.76 % and 0.01 % in three directions, indicating the one of analyzing workspace by improving the Monte Carlo method was effective. The research can provide the theoretical reference for orchard operation platform's structural parameter optimization, smart control and workspace analysis of similar type of agricultural machinery.