The ability of crop models to decompose complex traits and integrate the underlying processes enables them to capture genotype-environment interactions in diverse environments. Integrating genomics with biophysical crop models represents a potential breakthrough technology for improving our understanding of genotype-environment interactions across the biological organization. We present the results of a multi-model analysis on integrating crop modeling with genomic prediction. Seven rice models were evaluated on their ability to predict days to flowering in ten environments from parameters estimated through genome-wide association and genomic prediction, using a 13-fold cross-validation scheme. Phenotypic data were based on a rice diversity panel of 169 accessions with 700k markers. Significant associations with known flowering genes were identified for several model parameters. Although high accuracy was achieved for genomic prediction of model parameters in calibration, prediction accuracy was low for untested genotypes. We observed divergent model performance using genomic-predicted model parameters, which was attributed to photoperiod and temperature response curves, and number of calibrated model parameters. Several areas were identified for further research that could lead to better understanding the genetic control of complex trait formation and improved integration of genomics with crop modeling.