<p>For the first time, a novel method for the design optimization of a 20-story high-rise concrete structure, considering earthquake and wind loads, has been proposed. This approach employs a finite element (FE) model updating technique, developed through MATLAB programming leveraging the Open Application Programming Interface (OAPI) library within the ETABS software, enabling the application of advanced optimization techniques to identify optimal solutions that meet objective functions and design constraint conditions. To further enhance computational efficiency, a newly developed Proposed Surrogate Model (PSM) is introduced, capable of predicting the behavior of the high-rise structure without requiring direct analytical computations. The PSM is built upon a deep neural network (DNN), with its architectural parameters—such as the number of layers, neurons per layer, learning parameters, learning rate adjustment factors, and momentum coefficients—optimized using advanced optimizers. The efficacy of this surrogate model is validated by comparing its predictions with those from various DNN variants, including DNN with gradient descent (DNN-GD), DNN with gradient descent and momentum (DNN-GDM), DNN with adaptive learning rate (DNN-GA), as well as traditional machine learning models such as support vector machine (SVM) and K-Nearest Neighbors (KNN). Additionally, a sequential design approach integrating the PSM with the FE model updating technique has been proposed as an innovative optimal design method, offering significant time savings while maintaining high accuracy and adhering to design constraint requirements. The results of this study indicate that the sequential design method achieves efficiency and accuracy comparable to conventional FE model updating methods, with a marked improvement in execution time, and outperforms SVM, KNN and the variants of the DNN model in terms of predictive precision, solidifying its superiority across diverse structural optimization scenarios.</p>