Mixing time is usually used as an important parameter for quantifying the mixing performance of multiphase flow in a metallurgical bath. However, the topic of the mixing time of a Kaldo furnace for copper anode slime treatment has received little attention from scholars to date; this concept is important for guiding the optimization of high-efficiency smelting in the Kaldo furnace. In the present work, the mixing behaviour of gas-liquid two-phase flow in a Kaldo furnace was simulated using a scaled-down 1:4.5 water model fabricated from transparent acrylic plastic. A visualization technique and a quantitative characterization method for mixing performance in the Kaldo furnace—acid-base neutralization decolourization combined with RGB-based image analysis—were innovatively proposed. This approach can not only determine the global mixing level η¯(t) and track the frequency distribution of the local mixing performance ξk(t) of the fluid region during the mixing process but also quantify the mixing time. In addition, based on the proposed method and by using dimensional analysis, the effects of operating parameters (such as the gas flow rate Q, injection lance height h, rotation speed ω, liquid volume fraction φ, inclination angle θ, and liquid viscosity μl) on the mixing time of gas-liquid two-phase flow in the molten bath were discussed. On this basis, the empirical equation of the mixing time as a function of h, ω, φ, the modified Froude number (Fr'), μl and θ was established as the first quantitative relationship of the mixing time for the Kaldo furnace.