Protein structure prediction is essential for understanding biological functions, advancing disease diagnosis, and facilitating drug development. While experimental methods like X-ray crystallography, NMR, and cryo-EM provide precise structural data, they are constrained by high costs and low throughput, leaving much of the human proteome unresolved. Computational approaches have emerged as effective alternatives. This chapter offers a concise overview of protein structure prediction, beginning with the definition of protein structure and a brief discussion of secondary structure prediction. The exploration of diverse methods for predicting tertiary structures follows, including template-based homology modeling, threading using partial sequence templates, fragment assembly leveraging known protein fragments, ab initio folding based on molecular dynamics simulations, contact prediction informed by amino acid coevolution, and end-to-end prediction models. Particular emphasis is placed on AlphaFold2, a deep learning-based approach that has achieved unprecedented accuracy, revolutionizing the field. The transformative impact of such cutting-edge algorithms highlights the potential of computational methods to overcome limitations in experimental techniques, offering new insights into structural biology and paving the way for advances in understanding protein function and drug discovery.