The semi-airborne transient electromagnetic method (TEM) has been widely used in resource exploration and other fields due to its high efficiency, strong adaptability to the complex environment and high resolution. However, the conventional semi-airborne electromagnetic method continues to use the survey area of the grounded-wire TEM method. Compared with the existing ground-based TEM method, the semi-airborne TEM method is more sensitive to signal intensity because observed data is easily affected by noise for receiving position in the air. Therefore, it's unreasonable to consider only signal intensity or sensitivity related to offset in survey area. To obtain more suitable area, we discuss the distribution of the transient electromagnetic field with electrical source as well as the attenuation characteristics of the signal at different flight altitude through the forward modeling. Then, we obtain the sensitive area and flight altitude applied by the error analysis method in typical models. Finally, we describe a method what is herein dubbed the fuzzy comprehensive evaluation, which greatly reflects the membership degree of the evaluation object from different single factors to each level of fuzzy subset. Therefore, we use this method to get the optimal survey area and flight altitude under different noise levels. In low noise level, the signal intensity and sensitivity account for 33.3% and 66.7% respectively in selecting the optimum area and altitude, resulting in that the optimal area and flight altitude are coincides with the sensitive area and altitude better. While in the area filled with noise, the most area of strongest signal intensity is considered the optimal area and altitude. A field example shows that the aforementioned method can be well resolved in identifying survey lines and improving data quality.