AbstractGeneration of successful software project is the fundamental problem for many organizations. Therefore, the success of software project effort estimation is considered to be one of the foremost con-cerns of software engineering. Due to complicated nature of software projects, effort estimation has be-come a demanding issue that has to be addressed at the initial project stage. Here, Ensemble Classifier with Optimized Satin Bowerbird (ECOPB) for software project effort estimation is proposed. The ECOPB uses coherent neural network based clusters which are based on the deterministic model by updating mul-tiple centroids. Subsequently, an ensemble of Coherent Bayes with Selective Gradient Boosting Classifier is trained on the clustered result sets to build the classifier model based on projects’ attributes. Finally, Optimized Satin Bowerbird Effort Estimation is employed to optimize the classifiers. The proposed ECOPB is implemented using Cocomo81,CocomoNasa93, Cocomonasa60,desharnais,Albrechtfpa, ChinaFPA and cocomo_sdr datasets extracted from PROMISE software engineering repository. The ob-tained results are compared to state-of-the-art software project effort estimation approaches.