The discovery potential for detecting new heavy Majorana and Dirac neutrinos at some recently proposed high energy e + e − colliders is discussed. These new particles are suggested by grand unified theories and superstring-inspired models. For these models the production of a single heavy neutrino is shown to be more relevant than pair production when comparing cross sections and neutrino mass ranges. The process e + e − −→ ν e ± W ∓ is calculated including on-shell and off-shell heavy neutrino effects. We present a detailed study of cross sections and distributions that shows a clear separation between the signal and standard model contributions, even after including hadronization effects. The recent Super-Kamiokande results [1] provided a strong evidence for neutrino oscillations and non zero neutrino masses. This has motivated many theoretical models that imply new heavy neutrino states. These new particles are present in several grand unified extentions of the standard model such as SO (10) or E 6 . Many other new states are present in these models but neutrinos are expected to play a fundamental role in any consistent extended model. This comes mainly from the possibility that light neutrino masses can be connected to the Fermi and grand unified scales through m ν = v 2 F ermi /v GUT . This relation can be obtained from the "see-saw" mechanism with new heavy neutrino states. A fundamental point to be experimentally clarified is the Dirac or Majorana nature of neutrinos.The new heavy neutrino masses are experimentally bounded to be greater than 80-100 GeV [2,3] and the mixing with light neutrinos is expected to be small, even if there is some model dependence on these results. This means that these new neutral leptons could be detected only at the next generation of high energy colliders NLC at SLAC and TESLA at DESY. In this paper we turn our attention to these new possible lepton colliders [4]. New linear e + e − high energy colliders have been proposed, with center of mass energy from 500 GeV up to a few TeV. More recently µ + µ − and e − e − options were also proposed, as well as the electron-muon colliders.Some time ago it was noticed [5,6] that for some models the single heavy lepton production in e + e − −→ νN is higher than pair production e + e − −→ N N . Two main factors contribute for this difference. The first one is the mixing angle (s i ≡ sin θ mix ) single power in the lightto-heavy neutrino vertex, contrary to the double mixing angle power in the heavy-to-heavy neutrino vertex. The second factor is phase space suppression. If we suppose that all mixing angles are of the same order, then we have in Table I the vertices for heavy neutrino interactions in three different models [5] for new heavy neutrino states: vector singlet (VSM), vector doublet (VDM) and fermion-mirror-fermion models (FMFM). We call attention to the suppression factor for the ZN N vertex in the vector singlet model, which is not present for the other models. Throughout this paper we will suppose that mixing angles for heavy...