The agricultural machinery positioning, dynamic tracking and working area calculation in greenhouse were focused on, and an optimum indoor positioning algorithm was proposed based on multi-source data fusion theory. The above method firstly relied on inertia navigation measurement technology to estimate initial value of target positioning; then wireless RSSI ranging technology and weighted centroid algorithm were adopted to obtain positioning measurements; and optimum calculation was conducted with Kalman filtering algorithm to eliminate errors in the end, which were common in present measurement using single measurement technology and caused by data drift, signal obstruction, electromagnetic interference and other factors, in order to acquire accurate positioning information, realize real-time dynamic tracking and calculate working area. In Matlab simulation analysis, the positioning algorithm evaluation indexes were established to assess positioning effect and it was found that the proposed algorithm had better positioning accuracy and stability than single wireless RSSI indoor positioning algorithm. In the field test, the results showed that indoor positioning accuracy was not more than 0.125m and error was less than 0.4%, which could effectively meet demands of agricultural machinery positioning and real-time dynamic tracking in greenhouse.