This paper presents a novel implementation of neural networks that incorporates neuroplasticity principles inspired by biological neural systems and neurons. We introduce a simple adaptive neural network architecture that combines traditional backpropagation with Hebbian learning, connection pruning, and dynamic weight requalification. Our proposed Neuroplastic Neural Network (NPNN) incorporates mechanisms for structural plasticity through selective pruning of inactive connections and adaptive learning rates based on performance metrics. Experimental results on the MNIST dataset demonstrate that our NPNN model consistently achieves superior accuracy along with a 95.7% reduction in final loss, faster convergence, and a 16.67% improvement in inference time compared to standard neural networks. We observe that the integration of neuroplasticity mechanisms enhances the model’s ability to adapt to task demands and avoid local minima during training. This work hopes to contribute to the ongoing effort to develop more biologically plausible artificial neural networks with self-optimizing capabilities.