Light water reactors (LWRs) are the world’s dominant nuclear reactor system. Uranium (U)-fuelled LWRs produce long-lived transuranic (TRU) isotopes. TRUs can be recycled in LWRs or fast reactors. The thermal neutron spectrum in LWRs is less suitable for burning TRUs as this causes a build-up of TRU isotopes with low fission probability. This increases the fissile feed requirements, which tends to result in a positive void coefficient (VC) and hence the reactor is unsafe to operate. Use of reduced-moderation LWRs can improve TRU transmutation performance, but the VC is still severely limiting for these designs. Reduced-moderation pressurized water reactors (RMPWRs) and boiling water reactors (RBWRs) are considered in this study. Using thorium (Th) instead of U as the fertile fuel component can greatly improve the VC. However, Th-based transmutation is a much less developed technology than U-based transmutation. In this thesis, the feasibility and fuel cycle performance of full TRU recycle in Th-fuelled RMPWRs and RBWRs are evaluated. Neutronic performance is greatly improved by spatial separation of TRU and 233-6U, primarily implemented here using heterogeneous RMPWR and RBWR assembly designs. In a RMPWR, the water to fuel ratio must be reduced to around 50% of the normal value to allow full actinide recycle. If implemented by retrofitting an existing PWR, steady-state thermal-hydraulic constraints can still be satisfied. However, in a large break loss-of-coolant accident, the emergency core cooling system may not be able to provide water to the core quickly enough to prevent fuel cladding failure. A discharge burn-up of ~40 GWd/t is possible in RMPWRs. Reactivity control is a challenge due to the reduced worth of neutron absorbers in the hard neutron spectrum, and their detrimental effect on the VC, especially when diluted, as for soluble boron. Control rods are instead used to control the core. It appears possible to achieve adequate power peaking, shutdown margin and rod-ejection accident response. In RBWRs, it appears neutronically feasible to achieve very high burn-ups (~120 GWd/t) but the maximum achievable incineration rate is less than in RMPWRs. The reprocessing and fuel fabrication requirements of RBWRs are less than RMPWRs but more than fast reactors. A two-stage TRU burning cycle, where the first stage is Th-Pu MOX in a conventional PWR feeding a second stage continuous burn in a RBWR, is technically reasonable. It is possible to limit the core area to that of an ABWR with acceptable thermal-hydraulic performance. In this case, it appears that RBWRs are of similar cost to inert matrix incineration in LWRs, and lower cost than RMPWRs and Th- and U-based fast reactor recycle schemes.