In a complex and uncertain environment

the precision and reliability of robot localization using a single sensor is poor and easy to be interfered.Aiming at this problem

the observations information from laser telemeter and ultrasonic wave sensor is fused by using square-root unscented Kalman filter（SR-UKF）.According to the updated values of the states and error variances

the robots Monte Carlo location（MCL）density function of importance is constructed

and the two sensor respective advantages is combined by making full use of the redundant information collected by various sensors.The simulation experiments show that the robot Monte Carlo Location decisions based on multi-sensor fusion have large improvement in positioning precision and robustness

and the feasibility of the method is proved.