To address the issue of insufficient learning of features between different organs in the segmentation of pheochromocytoma images from partially annotated abdominal datasets， which leads to difficulties in distinguishing tumor and surrounding organ boundaries accurately， a Multi-scale and spatial Frequency feature-based image segmentation Network for pheochromocytoma （MF-Net） was proposed. Firstly， a Multi-scale Spatial Frequency Channel Attention module （MSFCA） was constructed， which enhanced the capture of inter-organ texture and boundary features by weighting and fusing the frequency domain information of image and the multi-scale feature maps from adjacent encoders， thereby highlighting the feature representation ability of tumor regions. Then， an Upsampling Multi-Scale Feature Fusion module （UMFF） was introduced， which combined upsampled feature maps of different scales to enhance the model's ability to recognize objects of varying sizes in image. Finally， an Adaptive Objective loss function （AOb） was utilized， which calculated the loss for the annotated abdominal organ labels and adjusted the loss weights according to the annotated organ categories， thereby optimizing learning process of the segmentation network. Experimental results show that on the abdominal organ and pheochromocytoma datasets， MF-Net has the segmentation accuracy improved by 3.33 and 3.18 percentage points， respectively， compared to the separately trained nnU-Net （no new U-Net）， with Dice similarity coefficient （Dice） and Normalized Surface Dice （NSD） reached 89.07% and 92.85%， respectively. On the external datasets， MF-Net has the Dice and NSD of 84.66% and 90.55%， respectively. In addition visualization results indicate that MF-Net can handle complex backgrounds and blurred boundaries in pheochromocytoma images better， providing better technical support for accurate diagnosis and treatment of pheochromocytoma.