针对直升机战场弹伤形态快速预测与抢修需求，基于Johnson-Cook本构模型和冲击有限元模型计算得到穿甲弹打击直升机铝合金机身的弹伤形态数据库，建立体现非对称弹伤特性的双椭圆函数模型，构建非对称弹伤形态与入射角度的关联函数，提出机身弹伤形态快速预测方法，并与有限元数值方法、基于遗传算法优化的支持向量机（GA-SVR）模型、实验结果对比验证。随后研究弹体口径、速度、入射角度与弹伤形态的影响特性。结果表明：弹伤形态快速预测方法计算得到的弹伤形态与实验测量结果吻合较好，弹宽误差为9.68%，弹长误差为7.03%。相比于GA-SVR模型和有限元数值方法，弹伤形态快速预测方法计算得到的左、右半长、弹宽MAPE误差分别为2.87%、4.32%、0.62%和2.92%、2.67%、0.64%，偏心距MAE误差为0.554和0.35，但计算时间分别减少88%和97.7%，计算效率显著提高。相比于弹体口径和速度，入射角度对弹伤形态的影响更大，入射角度与左、右半长呈二项指数关系，与偏心距呈三次多项式关系。

For the rapid prediction and repair requirements of helicopter combat damage morphologies, a database for aluminum alloy body structures subjected to conical-cup-shaped armor-piercing bullets was established based on the Johnson-Cook constitutive model and impact finite element models. A double elliptic function model was developed to describe non-symmetric damage characteristics, incorporating key parameters such as bullet caliber, velocity, and incident angle to establish a relationship function for non-symmetric damage halves. A rapid prediction method for combat damage morphologies was proposed and compared against finite element numerical simulations, support vector machine mode optimized by genetic algorithm（GA-SVR） models, and experimental results. The method demonstrates good agreement with experimental measurements, achieving maximum errors of 9.68% in width and 7.03% in length. Compared to the GA-SVR model and finite element numerical simulations, the rapid prediction method shows average errors of 2.87%, 4.32%, and 0.62% for the left and right halves of perforation diameter, width, and offset distance （MAE error of 0.554）, respectively, with computation times reduced by 88% and 97.7%, significantly improving computational efficiency. Results indicate that, among the three parameters—bullet caliber, velocity, and incident angle—the latter has a greater impact on combat damage morphologies. The relationship between the incident angle and the left and right halves of perforation diameter is described by a two-term exponential function, while its relationship with offset distance is given by a three-term polynomial equation.