Over the past twenty years, organic-rich shales have emerged as valuable systems for potential oil and gas production. Shale reservoirs are usually low in porosity, ultralow in permeability, and contain hydrogen-rich organic materials, making traditional low-field 1H NMR relaxation methods difficult to apply. The mobilities of the different hydrocarbon components can differ dramatically, from producible methane to the solid kerogen (the biopolymer precursor to petroleum). These considerations imply a “paradigm shift” in petrophysical NMR, currently based on the knowledge that (i) NMR is lithology independent (responds only to the fluids, not the matrix) and (ii) the total signal amplitude is proportional to total “porosity”. Both statements are not valid for organic shales. An additional challenge is the presence of solid and highly viscous organic materials, such as kerogen and bitumen, which are usually not detectable with the standard Carr-Purcell-Meiboom-Gill (CPMG) method on low-field equipment, with the rapid T_2 signal decay of these components limiting access to decay information. The work presented within this thesis focusses on the development of NMR methods for improving the characterization of the various organic components in heterogeneous organic shales with potentially producible hydrocarbons. A combination of high-resolution solid-state 1H and 13C NMR spectroscopy methods and 1H NMR relaxation measurements provided insights into the chemical and structural evolution of kerogen during thermal maturation. Part of the project is aimed at the optimization of pulse sequences which are appropriate for the investigation of shales under static (non-spinning) conditions and low magnetic fields. Various multiple-pulse sequences were explored and combined to develop new methods for identifying and quantifying both the immobile and mobile components in shales. A further study was focused on the application of double-quantum (DQ) 1H NMR to shales for robust separation between solid and liquid signals, and results show that this technique offers a potential method for obtaining improved description of hydrogen-bearing components with limited mobility in shale rocks.