The possibility of coexistence of superconductivity (SC) and antiferromagnetic long range order (AFLRO) of the two-dimensional extended t − J model in the very underdoped region is studied by the variational Monte-Carlo (VMC) method. In addition to using previously studied wave functions, a recently proposed new wave function generated from the half-filled Mott insulator is used. For holedoped systems, the phase boundary between AFLRO and d−wave SC for the physical parameters, J/t = 0.3, t ′ /t = −0.3 and t ′′ /t = 0.2, is located near hole density δc = 0.06, and there is no coexistence. The phase transition is first-order between these two homogeneous phases at δc. 74.25.Ha Correlation between the d-wave SC and AFLRO is one of the critical issues in the physics of the hightemperature superconductivity (HTS) [1,2]. Early experimental results showed one of the common features of the HTS cuprates is the existence of AFLRO at temperature lower than the Néel temperature T N in the insulating perovskite parent compounds. When charge carriers (electrons or holes) are doped into the parent compounds, AFLRO is destroyed quickly and then SC appears. In most thermodynamic measurements, AFLRO does not coexist with SC [3]. However, this is still a controversial issue. Recent experiments such as neutronscattering and muon spin rotation show that the spin density wave (SDW) may compete, or coexist with SC under the external magnetic field [4,5,6]. Remarkably, elastic neutron scattering experiments for underdoped Y Ba 2 Cu 3 O x (x = 6.5 and 6.6, T c = 55K and 62.7K, respectively) show that the commensurate AFLRO develops around room temperature with a large correlation length ∼ 100Å and a small staggered magnetization m 0 ∼ 0.05µ B [7,8,9]. These results suggest that AFLRO may coexist with SC but the possibility of inhomogeneous phases is not completely ruled out.For the theoretical part, the two-dimensional (2D) t-J model is the first model proposed[10] to understand the physics of HTS. Anderson proposed the resonatingvalence-bond (RVB) theory for the model about one and a half decades ago. The theory is reexamined again [11] recently. The authors compared the prediction of the RVB theory with several experimental results and found the theory to have successfully explained the main features of cuprates. This so called "plain vanilla" theory did not consider the issue of AFLRO, which must be addressed at very low doping. From analytical and numerical studies of the t − J model, it was shown that at half-filling, the d−wave RVB state with AFLRO is a good trial wave function (TWF). In this case, SC correlation is zero because of the constraint of no-doubleoccupancy. Upon doping, the carriers become mobile and SC revives while AFLRO is quickly suppressed. However, if the doping density is still small, AFLRO will survive. Thus SC and AFLRO coexist in the very underdoped regime [12,13,14,15,16]. Exact diagonalization (ED) up to 26 sites show that both SC and AFLRO are enhanced by the external staggered field. This result also ...