本文提出一种采用Ｈｏｐｆｉｅｌｅ神经网络（Ｈｏｐｆｉｅｌｄ Ｎｅｉｒａｌ Ｎｅｔｗｏｒｋ简称ＨＮＮ）优化的图象重建算法。将图象重建问题转化为ＨＮＮ优化问题，取重建图象的峰值函数最小以及原始投影与再投影之间的误差平方和最小作为图象重建的优化目标，作为能量函数构造连续型ＨＮＮ模型，由ＨＮＮ能量函数极小化可得到重建问题的优化解。这种方法具有简单、计算量小、收敛快、便于并行计算等特点。对照ＡＲＴ算法，用计算机模拟产生的无噪声投影数据检验新算法，验证了新算法的优越性。 In this paper, an image reconstruction algorithm based on Hopfiele neural network (HNN) is proposed. The problem of image reconstruction is transformed into the HNN optimization problem. The minimization of the peak function of the reconstructed image and the minimization of the sum of squared errors between the original projection and the re-projection are taken as the optimization objectives of image reconstruction. The continuous HNN model is constructed as the energy function from Minimizing the HNN energy function gives an optimal solution to the reconstruction problem. This method is simple, small amount of computation, fast convergence, easy parallel computing and so on. Compared with the ART algorithm, the new algorithm is tested by using the noise-free projection data generated by the computer simulation to verify the superiority of the new algorithm.