In this study, precision agriculture management zones were delineated using yield data over four years from the combine harvester equipped with yield monitor and DGPS receiver. Relative yields measured during each year were interpolated to 4 m2 grid size using ordinary kriging. The resultant interpolated yield maps were averaged across years to create a map of the mean relative yield, which was then used for cluster analysis. The mean yield map of post-classification was processed by applying majority filtering with window sizes that were equivalent to the grid sizes of 12, 20, 28, 36, 44, 52 and 60 m. The scale effect of management zones was evaluated using relative variance reduction, test of significant differences of the means of yield zones, spatial fragmentation, and spatial agreement. The results showed that the post-classification majority filtering (PCMF) eliminated lots of isolated cells or patches caused by random variation while preserving yield means, high variance reduction, general yield patterns, and high spatial agreement. The zoned result can be used as yield goal map for preplant or in-season fertilizer recommendation in precision agriculture. In this study, precision agriculture management zones were delineated using yield data over four years from the combine harvester equipped with yield monitor and DGPS receiver. Relative yield measured during each year were interpolated to 4 m2 grid size using ordinary kriging. The resultant interpolated yield maps The average yield map of post-classification was processed by applying majority filtering with window sizes that were equivalent to the grid sizes of 12, 20 , 28, 36, 44, 52 and 60 m. The scale effect of management zones was evaluated using relative variance reduction, test of significant differences of the means of yield zones, spatial fragmentation, and spatial agreement. The results showed that the post- classification majority filtering (PCMF) eliminated lots of isolated cells or patches caused by random variation while preserving yield means, high variance reducti on, general yield patterns, and high spatial agreement. The zoned result can be used as yield goal map for preplant or in-season fertilizer recommendation in precision agriculture.