In the growing quantum computing landscape, addressing security and privacy issues is vital. Third-party quantum services may attract users due to cost and availability, but pose IP theft risks. To mitigate this, we propose an edge pruning obfuscation method and split iteration for Quantum Approximate Optimization Algorithm (QAOA). This approach secures IP, increases reconstruction difficulty, with limited performance degradation (up to 10%) and low overhead costs (less than 0.5X).