Fires in wildland-urban interfaces (WUIs) are resulting from intertwined physical processes at different scales: landscape, settlement, parcel, building and material(Vacca et al., 2020) and are causing growing damage worldwide in context of climate change and large urban sprawl. The significant damage caused by these fires requires effective reinforcement of the resistance of structures and parcels against exposure to fire. Recent methodologies (Vacca et al., 2020; Benichou et al., 2021; Maranghides et al., 2022) have emphasised the need to look at these objectives by considering the spatial relationships between fuels, exposures and building resistance, in the perspective of a fire propagating in WUI according to so-called “fire pathwaysâ€. At parcel scale the fire pathways often involve ornamental vegetation, that highly raises the damaging potential of the wildfire, this vegetation being at short distance to the structures and having size comparable to building size. Horizontal and vertical discontinuities in this vegetation do largely impact the exposure (Cohen, 2008) and installing such discontinuities is becoming part of the protection regulation against wildfires in different countries (Maranghides et al., 2022). While the impact of embers has been extensively studied in the Insurance Institute for Business and Home Safety (IBHS) facilities (Manzello and Suzuki, 2014; Suzuki and Manzello, 2020), the effect of the fuel discontinuities on the reduction of thermal attack when approaching the building has been poorly addressed. This study aims at simulating in laboratory conditions a moving fire front pushed by wind and propagating from the surface to a canopy of ornamental vegetation, with fuel discontinuities. The experimental setup is composed of a fire lit in a surface excelsior fuelbed, propagating to a nearby vertical ornamental structure (excelsior and cypress) exposed to controlled high wind exposure (up to 10 m/s). The work is completed by a comparison of numerical modelling between the WFDS and FDS models. The first steps of this study are shown here, namely on zero-wind surface fire propagation, showing a coherent sensitivity of the rate of spread to fuelbed width and fuelbed load, and showing the ability of WFDS and FDS to reasonably reproduce this rate of spread. The ability of FDS to propagate to an isolated vertical tree with the same modelling processes is also well established.