In view of the factors affecting profit maximization, such as the relation between photosynthetic rate and CO2 concentration in greenhouse, the plant market price and the cost of CO2 enrichment, a decision-making method for CO2 control quantity was presented to achieve the profit maximization. Taking lettuce planting as an example, a non-linear model was set up for forecasting plant growth rate by neural network, and a market price model was established for forecasting plant market price by polynomial fitting function. In terms of the above two models and the cost of CO2 enrichment, the optimal set-point of CO2 concentration was found out. According to the experimental data of lettuce growth and the seasonal changes of market prices from 2004～2006 offered by Ling Jiatang wholesale market, prediction models were established and finally information fusion was achieved, which provided a theory for the decision-making of CO2 control quantity in greenhouse control system.