Composite materials are increasingly used in the aerospace and offshore wind energy due to their impressive specific mechanical properties. Modelling the effect of environment on the degradation of composite materials remains a significant challenge. Phase Field (PF) fracture models are promising methods to address this challenge. PF methods can accurately capture arbitrary crack paths and evolving interface and couple with multiphysics phenomena. The PF method is a variational formulation of fracture. It is based on the classical Griffith energy balance. Cracking initiates when the strain energy of the solid reaches a critical energy release rate (i.e. fracture toughness). This work aims to explore the effect of microstructural fibre bridging, matrix toughness, interface strength and moisture contents upon the macroscopic behaviour. In this work, a numerical framework combining phase fracture model, moisture diffusion and hygroscopic expansion is proposed to predict the environment-assisted failure of composite materials. We simulated three-point notched beam bending (TPB) tests to predict the microscale crack growth and the fracture resistance of the composite materials [1-2]. An embedded cell model is used to reduce the computational cost. The full composite microstructure is represented in the damage process zone as an embedded cell, while the remaining material is modelled as an anisotropic solid. A continuous displacement field between the embedded cell and the homogenised region is ensured by sharing the nodes at their interface. We validate our model against the experimental results. The coupled phase field fracture method successfully captures moisture distribution, swelling of fibre and matrix, fibre-matrix debonding. This computational framework enables a virtual tool to investigate the role of the microstructure, material properties and environmental effects. This will lead to more efficient and rapid designs for enhancing the design of next-generation transport and offshore wind turbine blades.