The occurrence of recombination between different regions of a genome is of interest in medical genetics, evolutionary and other related aspects of biology for various reasons. In the population genetics context, recombination is a critical issue for analysing within-species variation or variation at the population level: by averaging out the genealogical histories over some part of a genome, recombination reduces the level of stochasticity and therefore play a practical role in evolutionary inference. Traditionally, the methods for analysing genome data and making inference assumed the absence of recombination. If this assumption is wrong, the resulting inference about the evolutionary history of gene sequences are misleading and this makes the effective detection, estimation and characterisation of the recombination rate a very important issue. The many different approaches so far proposed for estimating this parameter from sequence data have been broadly classified as parametric and non-parametric based. When the former is fully used (in its various forms) for analysing data, the derived results will usually come from characteristically heavy computational workloads and are highly efficient. However, the method suffers from the setback that it is only useful for small data sets. The later is generally easy to implement, but the plausible error that may be associated with estimates raises concern. We propose a flexible and computationally attractive approach for estimating the population recombination rate from a sample of sequence data. Simulation results showed that the method is well calibrated with very encouraging properties.