Double network (DN) hydrogels exhibit exceptional mechanical properties due to their distinctive cross-linking network structure, which facilitates progressive damage within the sacrificial network. This results in an inhomogeneous deformation field during the necking stage. To better understand the damage and toughening mechanisms of DN hydrogels, we developed a multi-particle tracking (MPT) method and successfully measured the inhomogeneous deformation field during this stage. Our findings reveal the coexistence of three distinct damage phases, with necking phenomena identified as a phase transition among these coexisting damage phases. To determine the stress-strain relationship in DN hydrogels, we developed a machine vision-based method for measuring the stress field. By integrating this method with MPT, we spatially combined the strain field and stress field of the sample in situ, effectively capturing the non-monotonic stress-strain relationship of DN hydrogels. Furthermore, to connect the microscopic damage behavior of the network with the macroscopic mechanical response, we proposed an anisotropic damage model for DN hydrogels based on a micro-sphere framework. The cross-linking structures of DN hydrogels are modeled as two networks with chains evenly distributed across the surface of a unit sphere. Internal damage within the sacrificial network arises from chain scission, considering the effects of damage cross-interaction, while anisotropic damage is quantitatively characterized by the distribution of the maximum initial stretch ratio of chains in all directions. The changes in the network structure are related to the damage model, establishing a physically meaningful relationship between the network parameters and the damage state of the sacrificial network. Our model of anisotropic damage evolution provides insight into the underlying mechanisms of necking phenomena and the high fracture toughness observed at the micro-scale.