As a fundamental cryptographic primitive, hash functions are widely used in various applications like digital signatures, blockchain, cloud storage etc. Most of the existing schemes are made for generalized purpose and not domain-specific. This paper present EduHash, a domain specific and parallel hash function optimized for more secure credential verification in blockchain based education system. By integrating a multi-round substitution-permutation network, tree based parallel compression and a keyed sponge construction, EduHash helps in accelerating the diffusion and enhances the avalanche effect without trading off with security. All these components helps in achieving resistance to collision, and domain-specific authentication. Experimental evaluation is done with 10,000 academic records showcasing strong avalanche behavior with 50.02% bit change, ideal bit distribution, high entropy, and statistical resistance against known cryptanalytic attacks. Comparative analysis shown EduHash performs competitively across key security metrics. The results suggest EduHash as a viable solution for scalable, tamper-resistant, and decentralized educational credential verification.