System-environment interactions are essential in determining charge transfer (CT) rates and mechanisms. We developed a computationally accessible method, suitable to simulate CT in flexible molecules (i.e., DNA) with hundreds of sites, where the system-environment interactions are explicitly treated with numerical noise modeling of time-dependent site energies and couplings. The properties of the noise are tunable, providing us a flexible tool to investigate the detailed effects of correlated thermal fluctuations on CT mechanisms. The noise is parametrizable by molecular simulation and quantum calculation results of specific molecular systems, giving us better molecular resolution in simulating the system environment interactions than sampling fluctuations from generic spectral density functions. The spatially correlated thermal fluctuations among different sites are naturally built-in in our method but are not readily incorporated using approximate spectral densities. Our method has quantitative accuracy in systems with small redox potential differences (> k(b)T). Specifically, we find that the temporal correlations of site energies are critical in determining the coherent incoherent transition, while the role of spatial correlations depends on the nature of the systems. In a system with repeated bridge units of the same chemistry, spatially correlated fluctuations enhance the charge delocalization and charge-transfer rates; however, in a system of units with different site energies, spatial correlations slow the fluctuations to bring units into degeneracy, in turn, slowing the charge-transfer rates. The spatial and temporal correlations of condensed phase medium fluctuations provide another source to control and tune the kinetics and dynamics of charge-transfer systems.