This technical note is concerned with the optimal linear estimation problem in the presence of multiple packet dropouts in both sensor-to-controller and controller-to-actuator channels. The considered stochastic packet dropouts are independent and Bernoulli distributed. Based on a linear packet dropout model with stochastic parameters, a linear-minimum-variance (LMV) filter is proposed using the orthogonality principle. A sufficient condition for convergence of the steady-state LMV filter under a dc controller output is also given. Simulation results show the effectiveness of the proposed filter.