Concerning the problem that the existing influence maximization algorithms on social networks are difficult to meet the requirements of propagation range, time cost and memory usage on large scale networks simultaneously, a heuristic algorithm of Mixed PageRank and Degree (MPRD) was proposed. Firstly, the idea of reverse PageRank was introduced for evaluating the influence of nodes based on PageRank. Secondly, a mixed index based on reverse PageRank and degree centrality was designed for evaluating final influence of nodes. Finally, the seed node set was selected by using the similarity method to filter out the node with serious overlapping influence. The experiments were conducted on six datasets and two propagation models. The experimental results show that the proposed MPRD is superior to the existing heuristic algorithms in term of propagation range, and is four or five orders of magnitude faster than greedy algorithm, and needs lower memory compared to Influence Maximization based on Martingale (IMM) algorithm based on reverse sampling. The proposed MPRD can achieve the balance of propagation range, time cost and memory usage in solving the problem of influence maximization on large scale networks.