Laparoscopic surgical instruments are difficult in object detection due to their large-scale changes. Feature Pyramid Network (FPN) can effectively solve the multi-scale object detection problem. However, there are still some problems in FPN that limit the full utilization of multi-scale features. By analyzing the design problem of FPN, we propose the Multi-Attention Enhanced Feature Pyramid Network (MAFPN). First, we replace the convolutional block with feature selection module (FSM) that combines channel attention and global attention, which selectively maintains important information and enhances the expressiveness of features at each scale. Second, the global contextual information is captured by the self-attentive augmented fusion module (AAFM), which enriches the high-level feature information in the FPN and enhances the feature fusion effect. Finally, we use Dynamic Convolution Decomposition (DCD) to alleviate the impact of upsampling while enhancing the feature expression ability. Experimental results on the laparoscopic surgical instrument detection dataset m2cai16-tools-locations indicate that when ResNet50 is used as the backbone network, MAFPN improves the baseline network by 1.8 percentage points, to 96.5AP.