<p>Additive manufacturing is increasingly employed for direct fabrication in dentistry, relying on product quality. Dimensional accuracy and surface roughness of resin material being crucial factors in producing high-quality dental devices. However, there is a scarcity of research for influence of additive process parametric combination on the characteristics of material highlighting the need for further investigation to address this gap. A design of experiment is created using Response Surface methodology making 32 optimal combinations of layer thickness, print angle, infill density, exposure duration and lift speed, to fabricate and evaluate characteristics of ASTM standard specimen. Backpropagation neural networks model is developed for performance evaluation. Firstly, model efficiency is optimized by varying ANN architecture comparing best R² and least RMSE as performance metrices. Initially, based on experiment, accomplished least value for roughness is 0.226 microns and maximum accuracy is 98.34%. Later, two objective model is optimized using non-dominated sorting genetic algorithm to maximize accuracy and minimize roughness resulting in Pareto front. Conclusive combination provides least roughness of 0.2064 microns and best accuracy of 98.01%. These results are further validated. The proposed approach can simultaneously improve the surface quality and accuracy of resin parts in dental devices fabrication.</p>