Aeroengine onboard adaptive model can be built by estimating the biases of measurable parameters, so the CA (constant acceleration) model was introduced to establish the measurable variables state-space equations and observation equations, and adaptive Kalman filter algorithm was employed to estimate the measurable variables directly. The bias of measurable variables can be obtained by subtracting standard model computing values from the estimated result of measurable variables. When using the standard Kalman filter algorithm to estimate the measurable parameters, the results will offset prominently because of the great system state-space model error. The principle of standard Kalman filter and the impact of model error to the filter estimate results were analyzed, and the technology of dynamically adjusting the weight of state prediction in the filter estimate results was introduced, then the single factor adaptive Kalman filter estimation principle and the recursion formula were presented, which was aimed to make the estimation more accurate. In order to reduce calculation cost, the sequence filter was applied separately to process different measured parameters. The algorithm and system model were verified using the simulated data. The calculation results show that the designed filter can converge rapidly, and the computing speed is satisfied. The estimate results are better than the standard filter’s. So the adaptive filter algorithm has better estimation precision and have some engineering value. 