This paper first constructs a SFA model of three-element input by incorporating the element of land in the SFA model of Cobb-Douglas function. Then, adopting the method of Maximum Likelihood Estimate (MLE) , the paper analyzes the tendency of urban productivity development and its influencing factors based on the data of 34 northeast cities at prefecture level and above between 2000 and 2013. The results show that the average urban productivity in three northeast provinces is gradually improved, going up from 29.8% in 2000 to 54.9% in 2013, while the overall productivity remains low. From high to low, the rankings in productivity of the three northeast provinces are Liaoning, Heilongjiang and Jilin, with a significant difference of 26.7% between Liaoning and Heilongjiang and another significant 11.0% between Heilongjiang and Jilin. Factors that impede the improvement of urban productivity include government intervention in economy, upgrading of industries and administrative hierarchy while the construction of infrastructure in cities significantly promotes urban productivity.