This chapter deals with the principles and the applications in food science of 2D T1-T2 and D-T2 correlation techniques in time domain NMR (TD-NMR) spectroscopy. These approaches have demonstrated their potential for the structural and dynamic characterization of foodstuffs. Because relaxation or diffusion measurements result in time-decaying signals, the data analysis involves Laplace inversion, known to be an ill-conditioned and ill-posed problem but coupled with efficient iterative algorithms still allowing for 2D reconstructions. As the food matrices are heterogeneous, multiphase, and rather unstable over time, 2D methods were first applied to help assign multicomponent relaxation times to specific molecules (mainly fat and water) in specific chemical environments, as well as for real-time monitoring of the changes they undergo during storage or particular types of processing such as cooking or freezing.