It is significantly important in practice, for both manual and GPS based autopilot operation of agricultural field machinery, to obtain optimal operation direction, route and scheduling strategies for fields of known boundary information by systematic planning and optimization, which help improving the agricultural operations' efficiency, quality and utilization rate of machinery, reducing operating consumption(such as fuel and time) and labor intensity,increasing land productivity and effective crop yields.In this paper, a well-designed complete coverage path planning algorithm was proposed and complimented for a combine harvester when it works in field of 2d arbitrary polygon boundary even with obstacles in it.The algorithm firstly changed the internal obstacle boundaries into part of the normally external boundary, turning complex boundary polygon with holes into simple one by adding pairs of seam edges between the internal and external boundaries.Then the simple boundary polygon was divided into two zones:one is the external zone filled with contour-parallel path and another is the internal one filled with direction-parallel path.Three different transition strategies for direction-parallel paths including traditional turning transition (TTT) with U, Ω and T turn mode, jump transition (JT) via the external zone and cross transition (CT) via the internal zone considering the different re-coverage characteristics of region in the harvest process, were discussed in detail in the third step.Fourthly,the minimum transition distance (MTD) was figured out between arbitrary two direction-parallel paths according the above three transition strategies and stored in the MTD matrix.Finally the order of all direction-parallel paths was formulated as Travel Sales Problem (TSP) and its solution was obtained using the greedy search algorithm based on the MTD matrix data.Several typical experiments based on real field boundaries exported from Google Earth show that the proposed algorithm with its procedure can deal with fields of different shapes even with holes, run robustly and efficiently, get the inside operation direction, route, sequence, turning strategies and headland turning area automatically in one time's execution, and notably improve the quality and efficiency of combine harvest operation process.Further jobs will be proceeded to deal with problems like refuel, grains unloading and team work.