The invention discloses a semi-analytic shadow area forecasting method suitable for an elliptic orbit. A shadow area range of any inclined elliptic orbit in an orbit period is predicted by use of a semi-analytic and semi-numerical method. The method comprises the following steps: first obtaining a possible existence range of the shadow area by use of the analytic method according to a relation of the sun and an orbit; and then roughly judging whether the shadow area exists, and classifying the shadow type; and last solving out the shadow area existence range by use of the numerical method in the finite iteration time according to different shadow types. The analytic method is used for replacing the numerical integration and iteration as far as possible in the process of estimating any orbit shadow area region, the computing quantity of the shadow area forecasting is lowered, and an efficient method is provided for realizing the star automatic shadow area forecasting of an earth satellite and deep space probe using solar energy electric propulsion as the power.