Based on the model of arable land pressure, this paper calculated the series of cultivated land pressure in Shandong province from 1990 to 2005. The trend of arable land pressure was analyzed by using exponential regression model, polynomial regression model and GM (1,1) model. The combination forecasting model of cultivated land pressure in Shandong was established by means of Particle Swarm Optimization (PSO). The results show that: (1)it is feasible to apply the PSO to solve weight of combination forecasting model.(2) the arable land area is decreasing and the land area per capita decreased from 1990 to 2008, so cultivated land pressure is increasing. The pressure index will be up to 1.8438, 1.9081, 1.9745, 2.0433, 2.1145 and 2.1881 from 2010 to 2015 year, so it is difficulty to keep the arable land resource security. Some measures were proposed for easing arable land pressure.