The dielectric properties of vegetation reflect the coupling between the electromagnetic and physical properties of the material. Understanding dielectric behavior is essential for remote sensing applications. The traditional approach of measuring broadleaf permittivity involves exerting pressure on stacked leaves, which not only requires a complicated system to measure the pressure but also introduces a degree of uncertainty by changing the internal leaf microstructure. We propose a measurement technique that avoids the aforementioned shortcomings by using a vacuum packaging machine to remove the air from a sample bag, which is the first contribution of this paper. The proposed technique is validated against corn, cotton, and soybean leaf permittivity measurements. The second contribution of this paper involves how temperature is treated when modeling permittivity. The Debye-Cole dual-dispersion model, which is influential in vegetation remote sensing applications, is almost exclusively applied at 22 °C. However, it is assumed that this model can be extrapolated to other temperatures. Our analysis indicates that this model may overestimate permittivity at 28 °C, with mean relative errors (MREs) of up to 22% and 37% for relative permittivity and dielectric loss. Thus, we adopt the structure of the model due to its soundness and modify the model parameters. Overall, we used four different methods, among which the optimal method reduced the MRE by no more than -2.5%.