In the production process of sintering, the consumption of solid fuel is about 70% of the total energy consumption. There exists a nonlinear relationship between technological parameters and the solid fuel consumption in the sintering process. For the purpose of optimizing production and reducing energy consumption, this paper proposes an improved BP neural network to find the correlations between technological parameters and the solid fuel consumption. This paper proposes an improved conjugate gradient algorithm which combines the conjugate gradient algorithm with the inexact line search route based on the generalized Curry principle. The proposed algorithm has global convergence, optimizes the learning steps using new line search rules and improves the convergence speed. The new algorithm is applied in the cost prediction of actual sintering production. Simulation results show that the algorithm has better convergence compared with the traditional conjugate gradient algorithms. The MSE of prediction is 0.0098 and the accuracy rate reaches 94.31%.