The present work aims to develop a numerical modelling and simplified testing approach to simulate high velocity impact test on a filament wound glass fibre reinforced epoxy composite (GFRE) pipe under internal pressure. The composite pipe produced by filament winding process with fibre orientation of [±55]16. Initially, a virtual design of experiments (DOE) of several configurations was performed on a ring section of a pipe to achieve similar stress-strain components produced at the surface of composite pipe under pressure. The tensile loading of the fibres at the surface of the pipe due to internal pressure can be reasonably simulated by compressing a shorter pipe section of similar dimensions. Hence, lateral compressive loading normal to the area of interest was found to be the best choice to achieve the required strain components at the surface of the composite ring. Subsequently, a progressive damage and failure model to predict the damage produced by high/low velocity impact was achieved by combining a pre-peak non-linear response model and a post-peak behaviour softening model. The damage model consists of intra-laminar model and inter-laminar model. Intra-laminar model consists of pre-peak non-linearity (Shapery theory), failure initiation Hashin, and post-peak strain softening (Crack-Band). Each lamina was modelled separately by SR4 shell elements and joined together using layer-by-layer technique through surface-based cohesive contact considering traction separation law. Two dynamic explicit steps were performed which were lateral compressive loading followed by impact loading at 120 J. The experimental validation consists of fabricating a lateral displacement fixture, compressing the composite ring, followed by gas-gun (high-velocity) impact test. The results of experimental work for compressed and non-compressed ring combined with impact loading were validated with the numerical modelling prediction. An agreement was observed between the damage produced through the experimental with the numerical model prediction in most cases.