Abstract: A new chemometric method,the genetic neural network (GA-ANN) was proposed to optimize the structure and parameters of BP-ANN (artificial neural network with back-propagation algorithm).The self-adaptive probability of genetic algorithm was applied to search for the node number in the hidden layer,the learning rate (η) and the momentum factor (α).Optimal matching of the triplicate was achieved without the necessity of large quantity of training samples.The drawbacks of BP-ANN were overcome by GA-ANN.The optimized GA-ANN was successfully applied to the simultaneous determination of lanthanum and cerium in steel by CPA-mS-spectrophotometry.The apparent molar absorptivity for La and Ce were found to be 5.03×104 and 7.26×104L·mol-1·cm-1 respectively.