To solve the problem of class-based grasping in multicategory tiled scenes, this paper adopts different feature fusion methods and proposes a joint optimization algorithm MC-GSNet (Multi-Class GraspNet) that fuses category semantics and grasping posture and an optimization algorithm MT-GSNet (Multi-Task GraspNet) that builds a multitask learning model. The improved methods explicitly incorporate category information, optimize the generation logic of grasp poses and enhance the algorithm’s adaptability and success rate in multi-category object planar scenes. Experimental results on the public dataset GraspNet-1Billion demonstrate that the proposed methods significantly improve task adaptability and grasping success rates in multi-category planar scenes. MC-GSNet and MT-GSNet achieve 32.6% and 43.9% average accuracy improvements in grasp detection, respectively; MT-GSNet exhibits superior adaptability to unseen objects due to its integration of segmentation features. The experimental results in the simulation environment show that the grasp successful rates (GSR) of MC-GSNet and MT-GSNet reached 88.3% and 95.0% respectively, which can meet the needs of actual engineering deployment.