Reduction of non-working time for machines in field improves economical and environmental sustainability. Creating predefined tracks and routes in field maps by computer simulation, known as path planning, can reach solutions that increase the working efficiency of a machine. Also it becomes more important with the growing adoption of controlled traffic farming practices and with the use of robots as farm machinery. A number of works have been developed for path planning to achieve this, but they were limited in optimization because of their aim to obtain full coverage in non-squared fields, always reaching small corners that are difficult and costly to be operated, also the area overlapped in the headlands of these corners is representative. In this work an algorithm was developed that creates a series of tracks for a given field and a working direction and, afterwards, it detects corners where groups of tracks shorter than a given length are excluded from the field coverage, creating an uncropped area. By simulating different working directions in the field, an optimization to obtain the smaller loss for uncropped area, overlapped area and a reduced number of machine maneuvers is found. Two case studies were applied in the algorithm, an optimized path proposes a loss of 0.253% of area avoiding 5.69% turns on unprofitable tracks for a sugar-cane operation, and for a corn sowing case study, overlap of  coverage in headlands were 3.2 times higher in corner areas in relation to the whole field. The algorithm showed to be capable of optimizing paths on fields of different shapes and sizes.