In this paper, a new algorithm for attitude estimation in maneuvering flight, utilizing a combination of inertial measuring unit (IMU) and altimeter information is presented. Attitude estimation using a single IMU is possible only for near cruise flights, however for non-cruise flights, very large errors are obtained. In this paper, attitude estimation error is stabilized using an integrated IMU and altimeter system. The altimeter output being affected by gravity and the specific forces projected into the vertical plane bears insufficient information regarding the attitude states. Being a function of the roll and pitch angle, the specific forces will be in error, due to errors estimation of the attitude angles. Subsequently the vehicle vertical acceleration, speed and attitude will be inaccurate. In addition, due to a weak observability between the altitude measurements and the attitude angles to be estimated. For this reason and having a better estimate of the attitudes, the nonlinear attitude equations are converted into linear space, which will be beneficial for the estimation algorithm. Finally, simulation results using linear and unscented Kalman filters are curried out. A Monte Carlo simulation reveals that the newly suggested linear filter has a better performance in comparison with the non-linear unscented Kalman filter.