Identification of sensitive areas is a key point in targeted observations. This paper studies the manner in which conditional nonlinear optimal perturbation (CNOP) is used to identify sensitive areas in typhoon targeted observation and proposes three schemes including horizontal projection, single energy projection, and vertically integrated energy schemes. The sensitive areas determined by these three schemes have been compared, their physical meanings have been analyzed, and their advantages and disadvantages have been discussed. The efficiencies of the sensitive areas have been examined by ideal forecast experiments. The application of these three schemes to six typhoons showed that the sensitive areas determined by the single energy projection scheme were similar to those determined by the vertically integrated energy scheme and differed significantly from those determined by the horizontal projection scheme. Physical analysis revealed that the sensitive areas determined by the two energy schemes reflect the effects of environment on typhoons, whereas those determined by the horizontal projection scheme reflect the effects of asymmetry structures in typhoons. The ideal forecast experiments showed that the sensitive areas determined by two energy schemes are more efficient than that determined by horizontal projection scheme in reducing the energy of forecast errors in the verification area and improving the typhoon track forecasts, with the vertically integrated energy scheme showing the highest efficiency. While in improving the typhoon intensity forecasts, three schemes are comparable. Generally, the vertically integrated energy scheme is recommended when using the CNOP method to determine sensitive areas in typhoon observations.