This study aims at investigating the optimization of laser cutting Ni-based alloy, which is an important aerospace material. Firstly, the quality of laser cutting Ni-based alloy under variation of process parameters is evaluated by orthogonal experiment. The optimized parameters are analyzed intuitively and evaluated as 80.175. After that, the feedback neural network is used to train and fit the cutting quality, and samples of 17 and 18 are predicted that their error percentage of cutting quality are 3.14% and 2.20%, respectively. Finally, the genetic algorithm is used to optimize the extreme value based on this neural network model. The optimal fitness value and strain value within the population range are obtained through crossover operation with probability of 0.4 and mutation operation with probability of 0.2. The iterative evolution process is 100 times, and the theoretical score is 89.076. The actual score obtained by the verification experiment is 89.150, and the error percentage is only 0.074. Compared with orthogonal analysis and optimization, this method can quickly find the best process parameters with only a few experimental samples.