This is an experimental test of MD simulations on the picosecond timescale. Tryptophan zwitterion in TIP3P water at 278°K was simulated using CHARMM22 forcefield with the excited-state Trp atomic charges from [Toptygin et al., J. Phys. Chem. B 2010, 114, 11323]. Six stable excited-state rotamers of Trp sidechain were found with the population density peaks near (χ1,χ2) = (67°,80°), (−170°,57°), (−65°,115°), (65°,-85°), (−165°,-112°), (−65°,-80°). Curved boundaries between the rotamers on the (χ1,χ2) map were drawn along the troughs of the population density. Population density distribution within the boundaries of one rotamer reaches equilibrium in less than 20ps; equilibration between different rotamers takes much longer. At t>20ps rotamer populations can be described by a system of six first-order homogeneous linear differential equations. The solution is a sum of six terms Vmnexp(-t/τn). Population decay of each rotamer is not monoexponential and τn is not a lifetime. The same set of τn applies to all rotamers, but a different set of Vmn corresponds to each rotamer. The rotamers have slightly different fluorescence emission spectra, therefore fluorescence intensity is a sum of six terms αn(ν)exp(-t/τn), where αn vary with the photon energy hν. We have determined τn and αn(ν) in the global analysis of spectrally- and time-resolved fluorescence data (time resolution 65ps FWHM). Only four exponential terms could be resolved from the experimental data in H2O at 5°C (τ1 = 4780ps, τ2 = 2500ps, τ3 = 867ps, τ4 = 411ps); according to MD simulations the fifth term (τ5 = 241ps) has a very small amplitude, and the sixth (τ6 = 22ps) is faster than the time resolution. For a precise agreement between the experimental and simulated values of τn it is necessary to lower all potential barriers between rotamers by 0.178kcal/mol. This shows that fluorescence spectroscopy can be used to fine-tune torsional parameters.