The visible and near-infrared bands of AVHRR radiometer on the NOAA satellites can be used to monitor the Arctic sea ice cover. It is well known that Earth's surface characters in the cloud-covered regions could not be monitored exactly by the visible and near-infrared remote sensing. Therefore, the cloud detection is a fundamental process for sea ice parameters inversion. For the Arctic regions, the cloud detection algorithm has to focus on discrimination of sea ice and cloud. In this paper, the cloud detecting experiments for Arctic regions were conducted using the visible data of NOAA-19 AVHRR. The statistical characteristics of the different objects were analyzed as well. The results show that the cloud detection algorithm suit for land and ocean discrimination could not be used directly here. This is due to the similar radiation properties of sea ice and cloud. In the Arctic, on the basis of a large number of statistic analyses, the discrimination algorithm were re-screened with the thresholds being determined and revised, and then the preliminary cloud-ice discrimination algorithm for Arctic regions is obtained. This study can provide a basis for sea ice parameters inversion in Polar region, and then serve the polar sea ice monitoring.