A method to measure the thermal diffusivity of solid samples as a function of temperature is presented. The measurement technique is based on the mirage effect and in its linear zero-crossing relation for the transverse deflection, whose slope directly gives the diffusivity of the material. A 3D theoretical model has been developed in order to include both the effects of the radiative and convective heat transfers between the sample and its surroundings, and the temperature dependence of the refractive index and thermal diffusivity of the gas. The model also incorporates the effects introduced by the mirage parameters (sizes of the pump and probe beams, and probe beam height). The samples studied are opaque and thermally thick, and the applicability of the method is restricted to materials with diffusivity ≳1 mm2/s. Two experimental mirage setups are presented, one with the sample being heated in an open environment, and the other with the sample heated within a furnace. In the first case the range of measurable temperatures goes from ambient to ∼500 K, whereas in the second the upper limit is ∼1000 K. A comparison of the experimental results obtained with this method with those from the literature on calibrated samples of pure nickel, pure cobalt, and an AISI-302 alloy of low thermal diffusivity, confirm the validity of the model and method proposed.