By using the Support Vector Regression(SVR) machine and the genetic algorithm(GA)

a model for predicting the lower calorific value of fuel coal has been established

and the parameters of SVR machine model being optimized through GA.In order to reduce the dependency of selected parameters upon the training samples

the cross-validation ideas have been introduced

the parameters with best generalization performance being chosen as the final ones

these optimal parameters are input into the SVR machine model

and the genetic algorithm-support vector regression(GS-SVR) model has been obtained.Through analysis of the test data for furnace-entering coal in one power plant

and the comparison with the convintional SVR model and BP-ANN model

the reliability and accuracy of GA-SVR model have been verified

the results show that the said model can accurately predict the calorific value of fuel coals.