Abstract
The present invention relates to a wireless sensor network and a data sensing method thereof. A prediction model is built according to sensed data. When the statistic value of the sensed data, such as average value is in the allowed range of a user, the average value of past data is reported to the user according to the prediction model for reducing sensing times and reducing power consumption. Otherwise, when the statistic value of the sensed data, such as average value is beyond the allowed range of the user, actual sensed data is reported to the user according to the prediction model. Besides, the prediction model can be dynamically updated according to sensed data subsequently received.