Soft-sensing has been widely used in the case where the key variables are difficult to measure or can be measured but with a high cost. The traditional soft-sensing model is open-loop without correction mechanism. If the working condition is changed, the soft-sensing model which forecasts the following key variables will be not correct. In order to fetch the accurate value, it is necessary to carry out online correction. In this paper, a semiclosed-loop framework (SLF) is proposed to establish a soft-sensing approach. SLF approach estimates the input variables in the next moment by use of the prediction model and calibrate the output variables by use of the compensation model. The experimental results show that the proposed method has better prediction accuracy and robustness than other open-loop models in the problems of forecasting the sunspot numbers and flue gas oxygen content (FGOC).