Every year in the UK, 8-13% babies are admitted to a neonatal unit and 2-3% require intensive and/or high dependency care. Neonatal care in the UK has been organised through 'Managed clinical networks' (MCN's) since 2004. In England, there are 182 neonatal units organised in 24 networks. The baby charity BLISS reported that over a six month period, neonatal units were shut to new admissions for an average of 24 days, and 10% units exceeded their intensive care capacity for more than 50 days. The National Audit Office reported that capacity and staffing problems at unit level continue to constrain neonatal services. The main aim of this paper is to develop a model to assist with performance measure and capacity planning of such a neonatal network. We develop a continuous-time Markov chain model with overflow to capture movement of patients into neonatal units of a network using loss network framework. We then derive expression for overflow and rejection probabilities to measure performance of the network based on decomposition approach. The model is applied to a neonatal network in North Central London. Decision on number of cot can be made for specific level of overflow and admission rejection probabilities at each neonatal unit of the network.