Root length density (RLD) is a critical feature in determining crops potential to uptake water and nutrients, but it is difficult to be measured. No standard method is currently available for assessing RLD in the soil. In this study, an in situ method used for other crops for studying root length density and distribution was tested for sugarcane (Saccharum spp.). This method involved root intersection counting (RIC) on a Rhodic Eutrudox profile using grids with 0.05 × 0.05 m and modeling RLD from RIC. The results were compared to a conventional soil core-sampled method (COR) (volume 0.00043 m3). At four dates of the cropping season in three tillage treatments (plowing soil, minimum tillage and direct planting), with eight soil depths divided in 0.1 m soil layer (between 0-0.6 and 1.6-1.8 m) and three horizontal distances from the row (0-0.23, 0.23-0.46 and 0.46-0.69 m), COR and RIC methods presented similar RLD results. A positive relationship between COR and RIC was found (R2 = 0.76). The RLD profiles considering the average of the three row distances per depth obtained using COR and RIC (mean of four dates and 12 replications) were close and did not differ at each depth of 0.1 m within a total depth of 0.6 m. Total RLD between 0 and 0.6 m was 7.300 and 7.100 m m-2 for COR and RIC respectively. For time consumption, the RIC method was tenfold less time-consuming than COR and RIC can be carried out in the field with no need to remove soil samples. The RLD distribution in depth and row distance (2-D variability) by RIC can be assessed in relation to the soil properties in the same soil profiles. The RIC method was suitable for studying these 2-D (depth and row distance in the soil profile) relationships between soil, tillage and root distribution in the field.