To move in unknown or uncertain environment, a mobile robot must collect informations from various sensors and use it to construct a representation of the external world. Ultrasonic sensor can provide range data for this purpose in a simple cost-effective way. However, conventional ultrasonic sensor systems for a mobile robot are not sufficient for environment recognition because of their large beam opening angle and specular reflection. This paper describes on environmental perception algorithm which can solve these problems in case using ultrasonic sensor. The algorithm consist of two parts. One is to solve beam opening angle problem by data fusion from multiple ultrasonic sensors. The other is to cope with specular reflection problem in wall line extract, which is using Hough Transform. Experiments to verify the validity of the proposed algorithm are carried out, and the results are provided at last part in this paper.