Abstract: For the change-point model with at most one change i>i/n）=f（i/n）+ε（i/n）, where f(t)= \begincases\alpha_1+\beta_1\left(t-t_0\right), & 0ε（1/n）,...,ε（n/n）are indenpendently and identically distributed, the distribution of the estimator of the change point proposed in this paper can be approaximated by the first type of the extremal distribution with the help of the theory of Gaussian process. The problem of testing and interval estimation about the change-point to the magnitude of jump （α₂-α₁） and the magnitude of slope change （β₂-β₁） are considered.