In the devices like laptops, microprocessors, the electric circuits generate heat while performing work which necessitates the use of fins. In the present work, the heat transfer characteristics of hollow cylindrical pin fin array on a vertical rectangular base plate is studied using commercial CFD code ANSYS FLUENT© . The hollow cylindrical pin fins are arranged inline. The heat transfer augmentation is studied for different parameters such as inner radius, outer radius, height of the fins and number of pin fins. The base plate is supplied with a constant heat flux in the range of 20–500W. The base plate dimensions are kept constant. The base plate temperature is predicted using Artificial Neural Network (ANN) by training the network based on the results of numerical simulation. The trained ANN is used to analyse the fin in terms of enhanced heat transfer and weight reduction when compared to solid pin fin. Optimization of the hollow cylindrical pin fin parameters to obtain maximum heat transfer from the base plate is carried out using Genetic Algorithm (GA) applied on the trained neural network. The analysis using the numerical simulation and neural network shows that the hollow fins provide an increased heat transfer and a weight reduction of about 90% when compared to solid cylindrical pin fins.