Gross Regional Domestic Product (GRDP) data plays a vital role as a reference in regional development planning. However, the main challenge faced is the inaccuracy of GRDP growth predictions due to complex and fluctuating economic dynamics, especially in areas such as Simalungun Regency. Therefore, this study aims to analyze the development of Gross Regional Domestic Product (GRDP) by business field based on current prices in Simalungun Regency using three Artificial Neural Network (ANN) algorithms, namely Backpropagation, Bayesian Regulation, and Levenberg-Marquardt. The research data is GRDP times-series data for 2015-2023 obtained from the Central Statistics Agency of Simalungun Regency. The analysis used five models of the same architecture, namely 7-5-1, 7-10-1, and 7-15-1, with a target error of 0.01 and a maximum epoch of 1000 iterations. The results of the study indicate that the Levenberg-Marquardt algorithm with the 7-10-1 architecture model provides the best performance with an accuracy rate of 100% and the smallest Mean Squared Error (MSE) value of 0.0000214320 compared to other algorithms and architecture models. This finding indicates that the Levenberg-Marquardt algorithm is superior in predicting the development of GRDP in Simalungun Regency. The implementation of the results of this study is expected to help local governments and related agencies provide information on the development of GRDP in Simalungun Regency so that they can design more accurate and effective economic policies. In addition, this study also contributes to the development of artificial intelligence-based economic prediction methods, especially in the application of JST for the analysis of complex and dynamic regional economic data.