Now we extend one method into a sequence of binomial data, propose a stepwise confidence interval method for toxicity study and can identify a minimum effective dose. The first one is based on the well-known conditional confidence intervals for odds ratio, and the other one comes from Santner (1993)." small-sample confidence intervals for the difference of two success probabilities", and it produces exact intervals, through employing our method, all the declarations can be guaranteed to be correct with a probability higher than 100(1- )%. That is, the error rate is properly controlled.