Precision Agriculture (PA) is in its first steps in Brazil citrus production. Variable rate fertilization based on soil grid sampling and yield maps has been tested in São Paulo orchards. In a long term study results showed potential on increasing fertilizer use efficiency and improving soil fertility management. Despite the good results, in some cases it is noticed that systematic methods of investigation (grid sampling and yield data) and prescription (standardized prescription equations) are not suited to recognize areas that need different approach regarding soil fertility management. Management zones (MZ) delineation can distinguish such areas and provide means for appropriate management strategies based on spatial data such as soil EC, texture, elevation and sequential yield maps. Although MZ have been intensely studied and methods for delineation are fairly well developed, validation is still needed in Brazilian citrus orchards.  Soil survey might provide enough data for MZ delineation in most crops. But, in citrus, historical yield data is a key factor in delineating MZ. As a perennial crop, variability might be due to management actions (not necessarily related to soil) carried years ago that might still play important role in yield spatial variability. At the same time that yield data is an important layer in MZ delineation, in most cases it is not available since harvest is still predominately manual, and appropriate method must be applied to develop yield maps in such harvest type. The current yield mapping method adopted in our experimental orchards is based on georreferencing bags that are filled with fruit and distributed in the field during harvest. Yield is calculated based on distribution of bags in the orchard. Despite its simplicity this methods is efficient in providing yield maps. Also it is quite inexpensive and practical. Regarding methods for MZ delineation, cluster analysis is probably the most accepted. But, in some cases, if dedicated and user friendly software is not available the statistical steps of clustering is not the best option for PA practitioners. On the other hand, simple methods based on normalized historical yield data might provide reliable MZ in citrus, since yield can represent most of the spatial variability within an orchard. The objective of this study was to compare two methods of obtaining management zones in citrus orchards in Brazil, one based on cluster analysis of soil and yield data, and the other based on normalized historical yield data. Soil (EC, texture and organic matter) and yield data (from 2008 until 2011) were collected from two 25.7 ha orange orchards in São Paulo, Brazil. Method 1 was based on Principal Component Analysis and hierarchical clustering using ether soil and yield data. Method 2 was based on only normalized historical yield data. The yield average from four years and coefficient of variation (CV) was calculated in each pixel. Pixels were classified into two MZ, one of high yield (over 100%) and another of low yield (under 100%). Temporal consistency was assessed using CV (less than 10%). To verify MZ delineation in both methods, uniformity within each zone and disparity between zones were analyzed. In both methods, MZ delineation resulted as expected considering our experience and knowledge about these fields. In one of the fields texture gradient interferes in yield and in the other field a small portion of the orchard present drainage problems that also reflects in yield. It was noticed that this factors were well covered in the resulting MZ. Both methods resulted in similar MZ. Accurate comparison of the resulting MZ in the two methods is now being carried. So far, results indicate that simple method of MZ delineation based only on historical yield data is sufficient to provide reliable MZ in these fields. If historical yield data is available, growers can use this information to create MZs allowing better management strategies to their orchards.