The technology for the 20-MW(thermal) Transportable Fluoride Salt–Cooled High-Temperature Reactor (TFHR) is proposed by Massachusetts Institute of Technology for off-grid applications such as Antarctic bases and remote mining sites. The preliminary thermal-hydraulic analyses and improvements based on a 1/12th full-core model were performed using three-dimensional computational fluid dynamics (CFD). A benchmark study was conducted by comparing the CFD results against empirical correlations and experimental data obtained by Cooke, Silverman, and Grele. In the 1/12th full-core analysis, three practical considerations that may challenge the TFHR temperature limits are evaluated as bounding analysis. These include (1) helium gap between fuel compact and graphite block, (2) thermal conductivity degradations of graphite matrix due to neutron irradiation, and (3) full-core scale power distribution obtained from neutronic calculations. These design considerations lead to insufficient margin between the normal operating condition and the predefined thermal limits. In this context, additional design features are implemented to improve the thermal-hydraulic safety of the TFHR. First, bypass flow in the interstitial gaps between the active core and the reflector is found capable of reducing the temperature peaks at the core periphery. Second, improvements of the flow distribution from the central downcomer to individual coolant channels enable a higher mass flow rate to the regions with compromised cooling access. Overall, thermal-hydraulic performance was significantly improved with a fuel temperature margin from 10 to 150 K and a coolant temperature margin from 16 to 160 K, as well as the more uniform temperature distribution across the reactor core. Furthermore, thermal-hydraulic safety can be maintained at a 20% overpower operating condition [i.e., 24 MW(thermal)]. Overall, this study provides an engineering basis for the TFHR thermal-hydraulic design to improve its safety margin.