Quantum computers developed in recent years have tight resource constraints, and thus the resource usage reduction of quantum circuits has become a critical problem. Quantum state clean-up has high resource usage, and different quantum state clean-up strategies have different trade-offs between qubit usage and circuit depth. Previous work formulated the problem as the reversible pebbling game and developed a SAT-based algorithm that renders a valid quantum state clean-up strategy with qubit usage constraints. Our work improves the algorithm by formulating a new reversible pebbling game to handle the in-place operations directly. Experimental results show that our approach can improve qubit usage by 38.58% and ancilla qubit usage by 45.60% compared to the original approach.