Abstract: A method of omni-vision-based simultaneous localization and mapping(Omni-vSLAM) for mobile robots is introduced.The presented method extracts the color areas and uses the extracted information as the visual landmarks.The imaging principle and localization uncertainty of the omni-vision are analyzed to build the observation model and locate the landmarks.Then the robot poses and map information are updated synchronously with the help of extended Kalman filter(EKF).The experiment results show that the presented method can build the environment map and correct the accumulated location error caused by the odometer simultaneously.