Microsampling strategies were evaluated for the direct analysis of dried sugar cane leaves by energy-dispersive X-ray fluorescence spectrometry (EDXRF) and laserinduced breakdown spectroscopy (LIBS). The analysis by EDXRF was carried out by irradiating each leaf fragment in its middle portion with a collimated 5 mm X-ray spot size during 50 s, allowing the determination of P, K, Ca, S, Fe, Mn and Si. EDXRF was also useful to conclude that 15 leaf fragments (37.5 % of the recommended sampling area) were enough for attaining a representative analytical response from the whole diagnostic leaf. Regarding LIBS, that employs a substantial smaller ablation area (i.e., 750 µm laser spot size), sampling strategies were defined by taking into account the microchemical distribution of P, Ca, Mg, Fe, Mn, B and Si in 9 mm x 9 mm leaf fragment area. The proposed sampling protocol relied on the interrogation (rastering) of 3 equally spaced sampling lines in each leaf fragment with 48 accumulated laser pulses per line (Nd:YAG at 1064 nm, 5 ns, 10 Hz, 50 J cm -2 ) perpendicularly to the leaf midrib. This strategy enabled the simultaneous determination of P, K, Ca, Mg, Fe, Cu, Mn, Zn, B and Si by LIBS. Cross-validation between LIBS and EDXRF for P, K, Ca, Fe, Mn and Si predicted mass fractions presented high linear correlation coefficients of up to 0.9778 (selecting 15 leaf fragments per diagnostic leaf from 10 different sugar cane varieties). Results provide insights into a novel and promising strategy for direct and fast plant nutrition diagnosis, fostering further studies for in situ analysis of fresh leaves, strengthening the implementation of Precision Agriculture and Green Chemistry concepts. 3 % (for 15 fragments, 37.5 %diagnostic leaf area). The same trend was also observed for S, as the relative error diminished from 10 % (5 fragments, 12.5 % diagnostic leaf area) to 1.6 % (15 fragments, 37.5 % diagnostic leaf area). For the remaining evaluated elements, sampling 15 leaf fragments was appropriate for this intended purpose.Another alternative to evaluate the sampling representativeness was based on the construction of microchemical maps of P, Ca, Mg, Fe, Mn, B and Si in 9 mm x 9 mm leaf fragment area by LIBS for assessing the spatial distribution of the analytes. The selected number of accumulated laser pulses (i.e., 5 consecutive pulses) was defined as the threshold to perforate the leaf fragment at the experimental conditions set herein (50 J cm -2 , 750 µm spot size and 10 Hz). In addition, the microchemical maps were built