This paper proposes a pairwise SIR-based random sequential scheduling algorithm for wireless D2D communications. We derive an upper and a lower bound on the number of scheduled links by identifying the equivalence between the proposed algorithm and the Random Sequential Adsorption (RSA) process in physics. We then study the optimal SIR threshold, which is a key parameter in the proposed algorithm, for achieving the maximum sum rate. We finally extend the algorithm when a minimum SIR is required at each scheduled link. From the simulations, we observe that the proposed algorithm can achieve 24% higher sum rate compared with the aggregate SIR-based scheduling algorithm.