Aiming at lumber drying

characteristics of nonlinear and strong

coupling and hardly to build precise mathematical model

a modeling method based on wavelet neural network（ WNN） is proposed. Use datas collected by wood moisture sensor

temperature sensor and humidity sensor to build WNN model

and forecast measunement value of wood moisture sensor by model. WNN combines self-learning ability appropriating nonlinear problem of BP neutral network（ BPNN）

with representation ability for the local information of wavelet

so it has strong selfadaptive resolution and fault tolerance ability. Using collected datas in lumber drying process as training sample for simulation experiment. Result reveals that the model

which is built on WNN

can forecast measurement value of timber moisture content sensor

the WNN model has strong generalization capability and higher forecast precision than model built up by BPNN

which verifies feasibility and effectiveness of modeling of lumber drying kiln sensor by WNN.