Understanding rumen plant–microbe interactions is central for development of novel methodologies allowing improvements in ruminant nutrient use efficiency. This study investigated rumen bacterial colonization of fresh plant material and changes in plant chemistry over a period of 24 h period using three different fresh forages: Lolium perenne (perennial ryegrass; PRG), Lotus corniculatus (bird’s foot trefoil; BFT) and Trifolium pratense (red clover; RC). We show using 16S rRNA gene ion torrent sequencing that plant epiphytic populations present pre-incubation (0 h) were substantially different to those attached post incubations in the presence of rumen fluid on all forages. Thereafter primary and secondary colonization events were evident as defined by changes in relative abundances of attached bacteria and changes in plant chemistry, as assessed using Fourier transform infrared (FTIR) spectroscopy. For PRG colonization, primary colonization occurred for up to 4 h and secondary colonization from 4 h onward. The changes from primary to secondary colonization occurred significantly later with BFT and RC, with primary colonization being up to 6 h and secondary colonization post 6 h of incubation. Across all 3 forages the main colonizing bacteria present at all time points post-incubation were Prevotella, Pseudobutyrivibrio, Ruminococcus, Olsenella, Butyrivibrio, and Anaeroplasma (14.2, 5.4, 1.9, 2.7, 1.8, and 2.0% on average respectively), with Pseudobutyrivibrio and Anaeroplasma having a higher relative abundance during secondary colonization. Using CowPI, we predict differences between bacterial metabolic function during primary and secondary colonization. Specifically, our results infer an increase in carbohydrate metabolism in the bacteria attached during secondary colonization, irrespective of forage type. The CowPI data coupled with the FTIR plant chemistry data suggest that attached bacterial function is similar irrespective of forage type, with the main changes occurring between primary and secondary colonization. These data suggest that the sward composition of pasture may have major implications for the temporal availability of nutrients for animal.