The aim of this study was to investigate the extrusion cooking behaviours of two wheat flour samples (W1DK, W2CH) in two different twin-screw extrusion systems, where the wheat flour samples were characterised by rapid visco analyser (RVA) measurements. A new method is proposed to model and compare the wheat flour extrusion behaviours in the two systems through the RVA measurements, an extrudate bulk density model and a melt viscosity calculation. A new method is suggested to calculate the melt viscosity in the extrusion process. The average viscosity value of the W1DK sample was 64% higher than that of the W2CH sample in the RVA measurements. The comparison results show that the high viscosity for W1DK obtained from the RVA corresponds to a high reduced melt viscosity (η r , (1/s)γ) for W1DK, i.e., η r = 0.25–2.61 (1/s)γ (η r = 0.042–0.082 (1/s)γ for W2CH), a high specific mechanical energy (SME) for W1DK, i.e., SME = 8.16–11.28 kJ/kg (SME = 5.57–9.99 kJ/kg for W2CH) and a low extrudate bulk density (ρ B ) for W1DK, i.e., ρ B = 185–488 g/L (1 g/L = 1 kg/m3) (ρ B = 406–510 g/L for W2CH) in the two extrusion systems. The RVA viscosity and the calculated melt viscosity can be used to identify a recipe characteristic in terms of torque and extrudate bulk density in the two extrusion systems. A comparison has also been made for the bulk density model and response surface methodology in the extrudate bulk density prediction.