An exact maximum likelihood（EML）registration algorithm is presented

which is based on the effects of measurement noise.The registration estimates are obtained by maximizing the likelihood function.It shows that the EML criteria contains two sets of partially separable variables（the actual target positions and the sensor registration errors）and a two-step recursive optimization algorithm is provided.Through a computer simulation

it proves that EML has high accuracy and extensive in the process of multipe sensor registration and the solution to the limitation in minimum double-multiply registration is given.The given algorithm outmatches traditional minimum double-multiply registration especially when sensors are far apart from each other.