The European fusion roadmap [1] foresees the design and construction of the EU DEMO reactor to produce fusion electricity by 2050. The conceptual design of the machine is presently ongoing within the framework of the EUROfusion consortium. Some essential ingredients of the DEMO design come from that of ITER, the experimental reactor currently under construction at Cadarache, France. However, on one hand, the size and purpose of DEMO are more challenging than ITER and, on the other hand, research and development of the reactor sub-systems is achieving considerable improvements that make solutions feasible for the DEMO design, which were not for ITER. Superconducting magnets are one of those sub-systems: ITER and many design options for DEMO rely on the well-established technology based on the use of low temperature superconductors (LTS). However, recent improvements in current density, uniformity, and length of commercially available tapes, made by rare-earth barium-copper-oxide (REBCO) high temperature superconductor (HTS), make them potentially interesting for fusion applications. Different high-current cable-in-conduit-conductor (CICC) concepts are being proposed, cooled by forced-flow supercritical He, as the so-called CrossConductor (CroCo), recently designed at the Karlsruhe Institute of Technology [2]. The design of a large DEMO HTS magnet is however not straightforward, in view of both the lack of experience on large HTS magnets for fusion and the multi-scale thermal-hydraulic phenomena to be considered. While the former issue can be addressed by suitable simulation tools, the latter must be properly accounted for in these codes in order to perform reliable simulations driving the design choices. In particular, the spatial scales range from the thickness of a HTS tape (~0.1 mm) to the diameter of a CroCo macro-strand (few mm), to the CICC external size (few cm), up to a magnet height (~10 m). The time-scales involved range from that characterizing the He-to-strands convection (~10 ms) or the radial heat conduction in the strands (~10 ms), to the He transit time in a CICC (up to ~1000 s). Following the rationale already successfully adopted for the 4C code [3], one of the computational tools that has been employed since ~ 10 years for the design and analysis of LTS magnets, a multi-scale approach is being followed here to develop a new numerical tool for the thermal-hydraulic modeling of HTS magnets. A detailed full 3D model at the micro-scale is necessary to provide the constitutive relations, i.e. friction factor and heat transfer coefficients, to the meso-scale model. In the latter, each CICC is addressed by means of a new quasi-3D modeling approach, proposed here in order to reliably compute the most relevant thermal-hydraulic variables, such as the maximum conductor temperature during a thermal transient. Finally, at the macro-scale level, the entire magnet and related cryogenics are considered, similarly to what is done for LTS. [1] F. Romanelli, et al., "Fusion Electricity - A Roadmap to the Realisation of Fusion Energy, EFDA, 2012. [2] M. J. Wolf, et al., "HTS CroCo: A Stacked HTS Conductor Optimized for High Currents and Long-Length Production", IEEE Transactions on Applied Superconductivity, Vol. 26, no. 2, 6400106, March 2016 [3] L. Savoldi, et al., "The 4C code for the cryogenic circuit conductor and coil modeling in ITER", Cryogenics, Vol. 50, pp. 167-76, 2010