Abstract: In order to achieve precise control of the greenhouse environment, a fuzzy neural network was used to predict the greenhouse environment factors in view of the large lagging characteristics of the greenhouse environment. By determining the network structure, membership function and fuzzy rules and other parameters of the fuzzy neural network, the greenhouse microclimate prediction system was modeled. The data processed by wavelet noise reduction was used to predict the environmental factors through the greenhouse microclimate prediction model. The model verification results showed that the correlation degree obtained by using the temperature, humidity and light intensity model has reached more than 95%, which has a good predictive effect and can provide an effective basis for subsequent greenhouse control decisions.

Key words: ANFIS, greenhouse environmental, prediction model, greenhouse microclimate, fuzzy inference system