the objective of this work was to determine the coefficients of repeatability of morphophysiological traits in elephant grass (pennisetum spp.) genotypes, from data obtained during six evaluation cycles. the minimum number of measurements and genetic parameters were estimated. the experiment design was a randomized complete block, in a split？plot arrangement, with four n levels (control, 30, 60, and 90 kg ha？1 per cut) and 16 pennisetum genotypes (11 interspecific hybrids and five cultivars). the cycles consisted of evaluations in 2010 (4/21, 7/19, and 9/28) and 2011 (1/6, 4/7, and 8/3). the coefficients of repeatability were of medium？high magnitude for all variables, indicating that there was regularity between repeated measures. for the variables forage mass, plant height, leaf length and width, stem diameter, leaf chlorosis and area index, three cycles of evaluation are enough to reach r2 of 90% by principal component analysis. for internode length, the minimal of seven evaluations is necessary to predict the true value of genotypes. the genetic parameters of the variables forage mass, leaf length and width, stem diameter and leaf chlorosis are of high magnitude, favoring the selection of superior pennisetum genotypes.