Mixed stands of alfalfa-grass and pure alfalfa were sampled in farmers' fields in 19 New York counties during May and June 2004 and 2005. In addition field experiments with a randomized block design were established near Dryden and Ithaca, New York. Treatments included pure alfalfa (Medicago sativa L.) and mixtures of alfalfa with reed canarygrass (Phalaris arundinacea L.), timothy (Phleum pratense L.) and orchardgrass (Dactylis glomerata L.). A total of 1027 pairs of alfalfa and grass samples (single samples for pure alfalfa stands) and plant measurements were taken. Samples were separated and chemically analyzed for neutral detergent fiber (NDF), enabling development of prediction equations for mixed-stand fields. For both experimental plots and farmers' fields, alfalfa height, the key predictor in the PEAQ system, was shown to be significantly dependent (P<0.0001) on the percent grass, grass height, and grass species. This suggests that the PEAQ system may not be suitable for prediction of alfalfa quality in mixed stands. Numerous variables were significant (P<0.05) in predicting alfalfa NDF in both pure alfalfa and mixed stands of alfalfa and grass; however inclusion of alfalfa height alone resulted in a correlation coefficient of 0.85. A model with alfalfa height and growing degree days base 41 (GDD41) increased the correlation to 0.88. For both experimental plots and farmers' fields the addition of alfalfa stage to the model increased the correlation coefficient by less than one percentage unit, suggesting that alfalfa stage is not a biologically significant parameter for predicting alfalfa NDF in New York fields. For predicting the NDF of a mixed stand the most significant parameters included percent grass, GDD41, Julian date, alfalfa height, grass species, and grass stage. A model that included Julian date, alfalfa height and percent grass resulted in a correlation coefficient of 0.93 across all sites for 2004 and 2005.