An innovative method based on compressed sensing (CS) is presented for synthesizing subarrayed linear arrays using as few subarrays as possible with unequal sizes and weights. According to the CS theory, the related synthesis can be transformed into a sparse signal recovery problem since the element weight vector is compressible and can be sparsely represented due to only subarray-level excitation is conducted. Specifically, the synthesis herein is formulated as a convex problem with minimization of the l(1)-norm, and the corresponding parameters including the number of subarrays and the subarray weights and sizes can be determined simultaneously via convex optimization to satisfy the prescribed focused and/or shaped beam patterns. The proposed method is very easy to implement and has good computational efficiency. Numerical experiments are carried out to validate the effectiveness and advantages of the proposed method.