Abstract
We propose a new parametrization on dark energy, which is particularly reasonable and controllable. We invoke this parametrization to analyze the current supernova (SN) data and obtain the constraints in the parameter space at different confidence levels. We also invoke this parametrization to analyze the future observational re- sults expected by SNAP. We use the constant-w models to generate 2366 simulated SN data, use our new parametrization to analyze them, and then obtain the contours in the parameter space as the representations of those models. When the contours are disjoint, those constant-w are distinguishable. We show that when the interval between the two constant-w models is 0.05, the constant-w models can be distinguished completely by the simulated 2366-SN data with our new parametrization at the 2-sigma confidence level. When the interval becomes 0.1, the constant-w models can be distinguished at the 5-sigma confidence level.