Reconstruction after a disaster is essential to the reconstruction of communities that have been ravaged by either man-made or natural disasters. Errors in data calibration persist even as issues and challenges with planning reconstruction techniques are becoming more widely acknowledged. This study examines methods and techniques for creating resilient societies and reconstructing them following such occurrences in order to get around this restriction. Adhocratic Architecture effectively improves the performance of the post-disaster reconstruction process. Initially assessment and data collection is used and to analyze the extent of damage, the avail-ability of resources. Then Fuzzy logic algorithm used for feature extraction. After that we use Probability Graphics Models (PGMs) and Bayesian networks for analyzing the gathered data. subsequently Weakly Supplied training methods for classifying the general tags then Utilizing the Adam-Enhanced Squeeze and When a disaster strikes, the Excitation High-Resolution Network (SEHRNET) improves the precision of classifying structural damages. For balanced decision-making, the Preference Ranking Organization Method of Enrichment Evaluation (TOPSIS-PROMTHEE) technique and the Order of Preference by Similarity to Ideal Solution technique are utilized. Lastly, we employ Multi-time Stepping Rolling with Model Predictive Control (MPC) as a feedback loop to monitor the process. The proposed model is addressed utilizing several comparison metrics such as Accuracy, F1-score, Precision, Acceleration.