With ever increasing pressure from collider physics and direct detection experiments, particle physics models of TeV scale dark matter are gaining more attention. In this work, we consider two realizations of the class of scalar portal dark matter scenarios — the inverse seesaw model and the inert doublet model. Observations by the Cherenkov Telescope Array (CTA) of very-high-energy γ rays from dark matter annihilation in the context of these models are simulated for the Draco and Sculptor dwarf spheroidal galaxies, and later analyzed using ctools. We study the potential of CTA for the 5σ detection of a dark matter annihilation signal.In the absence of a signal, we also derive the 2σ upper limits on the annihilation cross-section. We compare our projected CTA sensitivity against the projected sensitivity of the next generation of direct detection experiment, i.e. XENONnT. Although the limits from CTA are significantly improved compared with the previous generations of γ-ray experiments, they are still ∼2 orders of magnitude above the thermal relic cross-section for the considered targets. In the case of the inverse seesaw model, the constraint from the future direct detection experiment XENONnT is much weaker than the CTA sensitivity, whereas for the inert doublet model, XENONnT gives a bound an order of magnitude stronger compared to the CTA limits.