Background: The ectopic pregnancy(EP) patients requires the closely monitor. However, there is no international consensus through which method to select EP patients from the pregnancy of unknown location (PUL) patients. Objective: To summarize and review the protocols of screening patients with EP when being diagnosed with PUL. Search strategy: We searched MEDLINE, web of science and Embase from inception to May 2020. All the articles were dual-reviewed based on predetermined selection criteria. Selection criteria: Studies exploring PUL outcomes can be included. Data collection and analysis: Prediction results, final diagnosis, and expense depending on patient’s visits and examination were analyzed by R. version 3.6.3 and Revman version 5.4. Results: 29 studies were included. M6 model had the areas under the curve(AUC) of 0.944, the progesterone cut-offs an AUC of 0.725, and the M4 model an AUC of 0.871 respectively. When the sum of visits and examinations of the protocols increased from 3 to 5, the rate of lost to follow-up patients increased from 11.19% to 18.63%. The average production utility of progesterone cut-offs is 0.242, the M4 is 0.174, and the M6 is 0.157. Conclusions: The M6 model had the best performance to predict EP among the PUL patients. The progesterone cut-offs is the most cost-effective method to predict the final outcome of EP. Fundings: Project supported by National Natural Science Foundation of China (82073323) and the Joint Funds of the National Natural Science Foundation of China (U20A20368). Keywords: ectopic pregnancy, pregnancy of unknown location, predictive protocol，hCG，progesterone.