This paper investigates the use of movable antennas (MAs) to enhance the security of wireless communication in the presence of multiple eavesdroppers. The authors consider a scenario where Alice, the transmitter, has no knowledge of the instantaneous non-line-of-sight (NLoS) component of the wiretap channel, but only has statistical information about it. They derive a tight approximation for the secrecy outage probability by interpreting Rician fading as a special case of Nakagami fading and using the Laguerre series approximation. To minimize the secrecy outage probability, they jointly optimize the transmit beamforming vector and the positions of the MAs at Alice. The problem is highly non-convex due to the complex incomplete gamma function in the objective. To address this, they introduce a slack variable and place the objective in the constraint, then approximate the inverse of the incomplete gamma function as a linear model. This simplifies the problem, which is solved using an alternating projected gradient ascent (APGA) algorithm. Additionally, they propose a sub-optimal scheme where Alice uses zero-forcing (ZF) beamforming, which is further simplified using a projected gradient descent algorithm. Numerical simulations demonstrate that the proposed schemes achieve significant performance gains compared to conventional fixed-position antenna schemes.This paper investigates the use of movable antennas (MAs) to enhance the security of wireless communication in the presence of multiple eavesdroppers. The authors consider a scenario where Alice, the transmitter, has no knowledge of the instantaneous non-line-of-sight (NLoS) component of the wiretap channel, but only has statistical information about it. They derive a tight approximation for the secrecy outage probability by interpreting Rician fading as a special case of Nakagami fading and using the Laguerre series approximation. To minimize the secrecy outage probability, they jointly optimize the transmit beamforming vector and the positions of the MAs at Alice. The problem is highly non-convex due to the complex incomplete gamma function in the objective. To address this, they introduce a slack variable and place the objective in the constraint, then approximate the inverse of the incomplete gamma function as a linear model. This simplifies the problem, which is solved using an alternating projected gradient ascent (APGA) algorithm. Additionally, they propose a sub-optimal scheme where Alice uses zero-forcing (ZF) beamforming, which is further simplified using a projected gradient descent algorithm. Numerical simulations demonstrate that the proposed schemes achieve significant performance gains compared to conventional fixed-position antenna schemes.