The use of sensor in the machine’s metering device allows to know the exactly quantity of the product to be distributed. Based on this information, it is possible to visualize the real amount applied, known as as-applied maps. There is no point in all the effort expended to visualize the variability if the machine does not have the ability to reproduce it correctly. It is known that there is a difference between the recommended and the real application rate. The purpose of this work will be to quantify the difference between the recommended and the real rate during the application of potassium chloride, lime and phosphorus in the orange crop, identifying the causes and factors with the greatest influence for such errors. For the study, as-applied reports of several fields were used and compared to the recommended values. With the aid of a spreadsheet system, a filtering was performed on the data, deleting points that do not characterize the correct use of the machine. It was noticed, for the calibrated machine, that the main differences occurred during the acceleration / deceleration periods and variation in the amount applied. Positioning errors can also be sources of errors in as-applied maps. It was verified that the assertiveness of the machine on the potassium chloride application was within the tolerance of ± 10% of the recommended value above 95% of the working time in two fields analyzed in 2011. The differences occurred almost exclusively in acceleration and deceleration points due to the machine response time. The application of potassium chloride, lime and phosphorus during three years will still be analyzed. The difference between the real and recommended quantity is expected to follow the same pattern as that already found. As-applied maps, although not frequently used, are excellent tools for assessing the quality of the application.