Surface reconstruction with high efficiency and accuracy is required in advanced aerospace manufacturing and measurement. As wildly used models

B-spline and NURBS are the main tools for surface reconstruction in existing methods. However

the topological constrains of B-splines and NURBS control meshes make it produce a lot of redundant control points when modeling surfaces. Newly proposed T-splines in recent years have provided a method to solve this problem. Analysis-suitable T-splines (ASTS)

a subset of T-splines with certain topological constrains were later developed

which optimized the mathematical properties and localized refinement capability of T-splines. T-splines have been successfully applied to surface reconstruction

while ASTS-based surface reconstruction algorithms are in vacancy at the moment. This paper proposes a surface reconstruction method based on analysis-suitable T-splines. Adaptive construction of an initial AST-mesh and efficient local refinement are performed in the algorithm to generate an ASTS surface within certain tolerance. Our experimental results show that the proposed ASTS fitting can generate spline surface with higher accuracy than NURBS fitting with the same amount of control parameters. The computing efficiency of the proposed algorithm is higher than simple T-spline fitting.