Quantum chemistry is a promising application in the era of quantum computing since the unique effects of quantum mechanics that take exponential growing resources to simulate classically are controllable on quantum computers. Fermionic degrees of freedom can be encoded efficiently onto qubits and allow for algorithms such as the Quantum Equation-of-Motion method to find the entire energy spectrum of a quantum system. In this paper, we propose the Reduced Quantum Equation-of-Motion method by reducing the dimensionality of its generalized eigenvalue equation, which results in half the measurements required compared to the Quantum Equation-of-Motion method, leading to speed up the algorithm and less noise accumulation on real devices. In particular, we analyse the performance of our method on two noise models and calculate the excitation energies of a bulk Silicon and Gallium Arsenide using our method on an IBM quantum processor. Our method is fully robust to the uniform depolarizing error and we demonstrate that the selection of suitable atomic orbital complexity could increase the robustness of our algorithm under real noise. We also find that taking the average of multiple experiments tends towards the correct energies due to the fluctuations around the exact values. Such noise resilience of our approach could be used on current quantum devices to solve quantum chemistry problems.