It is the primary step of "digital city" to construct three-dimensional digital map [1][2]. We consider that it constructed by computer vision based on the image microminiature aircraft collected [3]. But the characteristic point?s quantity cannot meet the requirement of construct Digital Elevation Model. This paper provides a latest correlative matching method based on Epipolar Constrain after introduce the characteristic point extract and matching. Also it adopted a precise Fundamental Matrix estimate algorithm to calculate epipolar equation and increase the matching point. This algorithm performs very well in terms of robustness to outliers and noises .It improves the precision effectively. It also gives the experimental evidence of this algorithm?s effect and its speed meets the real time 's request .