Abstract: Objective To observe the effects and safety of intra-thoracic or intraperitoneal administration of bevacizumab (Bev) in patients with malignant pleural effusion or ascites, and to explore the potential factors that influence the efficiency. Methods Clinical data of sixty-one cases with malignant pleural effusion or ascites who received intra-cavity injection of drugs were retrospectively analyzed, of which twenty-four patients were injected with bevacizumab-contained regimens, and thirty-seven patients were injected with traditional chemotherapy regimens. Paracentesis-free interval (PaFI) of the groups were compared to evaluate efficiency side events after administration of drugs were recorded. Cox regression was used to analyze the clinical characteristics, blood and general effusion test. Results Median PaFI in bevacizumab group was 115.0(55.3, 219.0) days, and was significantly longer than traditional therapy (53.0(32.0, 147.5) days), (Z=2.340, P=0.019). After stratified with effusion site, Kaplan-Meier analysis also demonstrated a significant benefit from bevacizumab therapy (χ²=7.568, P=0.006). Cox regression analysis concluded effusion site and ratio of monocytes and multinucleate cell counting in the effusion had influence on outcomes in addition to bevacizumab injection (all P<0.05). No differences were seen in side events between the two groups, side events rates were 41.67% for bevacizumab group versus 40.54% for traditional therapy group (χ²=0.008, P=0.930). The most common side event was catheter-related complications, not due to drugs. Conclusions Intra-thoracic or intraperitoneal administration of Bev was thought to more effective than traditional regimens in controlling malignant effusion with tolerable side effects. Effusion sites and ratio of monocytes and multinucleate cell counting in effusion may influence the outcomes.