Long linear array is more important than shipborne sonar in underwater target detecting and positioning, but it seems more sensitive to the shape distortion at the mean while. In this paper, the array shape distortion effect on soft linear array Direction of Arrival (DOA) estimation performance was studied theoretically, then with simulation and field data process, an array calibration method was put forward. The steering vector of beamforming algorithm was changed when array shape was distorted, which will cause the degradation of line array direction performance, so far as to failure. Based on the conventional line array algorithm: Conventional BeamForming (CBF) and Adaptive BeamForming (ABF), in the interest of process the data recorded from the distorted array, this paper first derived the general arbitrary array beamforming algorithm, and based on the impact analysis of array element position error act on the acoustical signal transmission delay, introduced how the beamforming steering vector was influenced by the array shape distortion. Next, an array shape calibration method based on acoustic signal was given, which correct the steering vector through estimating the array shape distortion. At last, the simulation and field data process proved the analysis, the performance of both CBF and ABF beamforming algorithm degradation which caused by the array shape distortion, in contrast, the CBF method was better, it has more robust than ABF method; The accuracy of DOA estimation output and the 3dB width of main lobe have been significantly improved after appropriate array shape distortion adjustment, for both CBF and ABF algorithm. The method mentioned here also useful to analyze the situation of three dimensional space