This paper proposes the use of a steady-state model to identify the parameters of an induction motor. A time-varying impedance can be found with the time-varying voltages and currents under a starting no-load low-voltage test, where time-varying slip rates were also recorded. With a proper conversion, the variations of impedance versus slip rates could be acquired as sampled data for identification, and both electrical and mechanical parameters can be found. A least mean square (LMS) method was used with a particle swarm optimization method to solve the aforementioned problem. The method to find a good set of initial values for LMS is also described in this paper. Many tests were conducted to simulate the starting phases of an induction machine to compare with the existing data. From various tests, the practicability and accuracy of this method can be proven.