This paper presents a novel method for mobile robot localization based on multi-robot cooperative sensing. A multi-robot cooperative localization scheme is proposed using visual detection and sensor data fusion. The visual detection system employs a stereo vision module for both observing other robots and obtaining environmental information. Each mobile robot is able to recognize its teammates by using the vision system. The localization error is reduced through the proposed sensor fusion algorithm. The cooperative localization algorithm consists of two stages: serial fusion and parallel fusion. Serial fusion aims to identify the positional uncertainty of an observed robot while parallel fusion reduces its positional error based on Kalman filtering. The multi-robot cooperative localization system has been realized through the server-client architecture. Experimental results are presented to validate the effectiveness of the proposed algorithms.