Localization in wireless sensor networks gets more and more important. This thesis presents novel localization algorithms based on RSSI for wireless sensor networks. Our approach is using mathematical model to estimate the position of sensor node. Although trilateration is a basic method in localization based on RSSI, the path-loss can be a source of major error. For getting more accuracy, we improve trilateration algorithm to reduce path-loss. Circle Intersection Points (CIP) and Circle Intersection Lines (CIL) can easily use RSSI to estimate position by only four reference nodes. The advantage of algorithm we propose is not only suitable for the rectangular area but also suitable for the irregular area. The experimental results presented in this thesis demonstrate the ability of this system to estimate position with a high degree of accuracy.