Summary: The maximum likelihood method is used to find the estimation of the unknown parameter for Burr XII distribution under type-I doubly censored sample. It is proved that the maximum likelihood estimation is uniquely existent one. The iterative formula and asymptotic variance of the unknown parameter are obtained by EM algorithm and it is illustrated by relevant lemmas that the iterative formula has good convergence. The random simulation result shows that the EM estimation of the unknown parameter is superior to the maximum likelihood estimation and as for type-I doubly censored sample, the EM algorithm is a better method for estimation of parameter.