Recently, Zhang et al. (2010) proposed an effective performance model for dwell-based eye pointing. However, their model was based on a specific circular target condition, without the ability to predict the performance of acquiring conventional rectangular targets. Thus, the applicability of such a model is limited. In this paper, we extend their one-dimensional model to two-dimensional (2D) target conditions. Carrying out two experiments, we have evaluated the abilities of different model candidates to find out the most appropriate one. The new index of difficulty we redefine for 2D eye pointing (IDeye) can properly reflect the asymmetrical impact of target width and height, which the later exceeds the former, and consequently the IDeyemodel can accurately predict the performance for 2D targets. Importantly, we also find that this asymmetry still holds for varying movement directions. According to the results of our study, we provide useful implications and recommendations for gaze-based interactions.