This thesis provides a methodology and set-up of a reactivity monitoring tool for Accelerator-Driven Systems (ADS). The reactivity monitoring tool should guarantee the operation of an ADS at a safe margin from criticality. Robustness is assured in different aspects of the monitoring tool: the choice of the measurement techniques, the evaluation methods to derive the reactivity from experimental data, and the detector type and positioning. In the first chapter of the work, the experience from previous research programmes (mainly MUSE and YALINA) is used to select appropriate experimental techniques for reactivity monitoring. A combination of three techniques is assessed to monitor reactivity: The Pulsed Neutron Source (PNS) technique (with the Integrated Source Jerk (ISJ) method as alternative) for start-up (until 1 % of nominal power) and the Current-to-Flux (CTF) combined with the Source Jerk (SJ) technique during ADS operation. Static evaluation methods are used to derive the reactivity from the experimental value: the area method for start-up, the static CTF method during operation, and the static SJ method for the interim cross-checking of the ADS reactivity. These are more robust than dynamic evaluation techniques, as they do not depend on the knowledge of kinetic parameters. In the second chapter of this thesis, the spatial correction factors (SCFs) are defined that need to be applied on the reactivity values obtained by point kinetics. Via modal analysis, the SCFs are derived analytically for the selected evaluation methods. Extensions are made to the existing DALTON diffusion code in order to perform modal analysis and time-dependent experiment simulations. In this way modal analysis helps to understand the behaviour of the simulated SCFs of the complete subcritical core. In the third chapter the proposed reactivity monitoring methods are tested on the zero-power VENUS-F ADS. Thanks to the modular core, VENUS-F can operate in both critical and subcritical mode. The experimental reactivity results can therefore be compared to a value obtained by an alternative methodology, starting from a critical core. The experimental results confirm the modal analysis and experiment simulations. For both absolute reactivity measurement techniques (PNS area method and SJ method) SCF < 1, i.e. a safe overestimation of the absolute reactivity value, is found outside a sphere at the centre of the core with a ?radius of about 20 to 25 cm, depending on the energy group. The SCF=1 location corresponds to the zero crossing of the first eigenfunction (different from the fundamental one) with a zero crossing outside the centre of the core. The outer reflector is chosen as robust location for the detectors during start-up and interim cross-checking reactivity monitoring during operation, as in this zone SCF < 1 with a small spatial gradient. For the VENUS-F first subcritical core SC1, SCF(area)=SCF(SJ)=0.96-0.97 (depending on the detector type), which corresponds ??to an overestimation of the absolute reactivity of 3709 pcm by 111-148 pcm. Special attention should be paid to (local) thermalising elements, considerably influencing the SCF. This issue could be solved by the use of threshold detectors. Finally, reactivity monitoring during operation in the MYRRHA ADS is studied in order to take into account the full-power aspects of reactivity monitoring. For the start-up reactivity monitoring and interim reactivity cross-checking MYRRHA behaves similar to VENUS concerning the SCFs for the evaluation methods. The (outer) reflector is therefore the recommended detector zone, preferably using threshold fission chambers as instrumentation. For the simulation of reactivity monitoring during operation, MCNP is used as appropriate simulation tool to deal with the MYRRHA core inhomogeneities. The relative CTF monitor is evaluated between three stages of operation: the temperature feedback during core start-up (from 1 % nominal power on), the core enlargement during the first MYRRHA cores, and the burn-up during the MYRRHA equilibrium core. Again, the MYRRHA outer reflector turns out to be a suitable detector location for reactivity monitoring also during operation, independent of the detector type. A safe (under)estimation of the absolute reactivity decrease is obtained for all stages of operation, except for the temperature feedback case. For practical reasons, only a limited number of detectors will be installed in the reflector. Therefore, a conservative boundary value of 500 pcm could be appropriate to take into account the locations with extreme SCFs, in case no re-calibration of the relative CTF monitor via the SJ method is possible. As a conclusion, it is possible to measure the reactivity in an ADS by combining different experimental techniques. The accuracy of the reactivity monitor will de- pend on the safety function it has to fulfill and the related licensing requirements. If it is possible to perform short (in the order of hundreds of ?s) reproducible accelerator beam interruptions with a stable beam restoration, the absolute re?activity during operation can be determined via the SJ cross checking technique. In that case, the accuracy of the reactivity monitoring tool can be improved. The higher the accuracy, the higher the operational keff level, which is beneficial for the operational cost of an ADS.