The paper presents a novel idea of proposing the application of Variational Autoencoders (VAEs) in crime detection for predicting face aging and deaging, which is one of the potential challenge of forensic science. VAEs are equipped to generate realistic aged or rejuvenated facial images in simulating the appearance of suspects over time. Forensic age progression through VAEs could be pivotal in matching facial images with existing databases, potentially advancing criminal investigations. This work has considered learning rate and mean square error (MSE) which play a crucial role in determining the quality and efficiency of the image reconstruction process. Structural Similarity Index (SSI) and face recognition accuracy parameters are considered to evaluate the proposed work and to compare the work with the existing literature. The precision and recall of the proposed method is 0.923 & 0.936 which is promising when compared with the existing literature.