Abstract: For the purpose of studying the goodness-of-fit testing of Logistic distribution, the authors discussed the titular problems. The results show that if the example is big enough and the type Ⅱ censored ratio is constant respectively, the approximate maximum likelihood estimation of distribution parameter still exists uniquely, has limit and is consistent to the real parameter; Particularly, if the cansored ratios tend to be a constant, the maximum likelihood estimation has the asymptotic normality.