Real-Time Job Scheduling is an important issue in the area of Real-Time systems. In many real-time systems, once a monitor job has started, another monitor job must be executed within a specific time interval (temporal distance) in order to monitor the progress of real-time processes. The problem (TDCP) of scheduling a set of jobs in which pairs of the jobs must be executed within a temporal distance constraint on a single processor is considered in this paper. For a set of jobs with arbitrary execution times, Han, Lin and Liu have already shown that TDCP is NP-hard in the ordinary sense even with just one temporal distance constraint. For a set of jobs with equal execution time, the problem will be related to the Linear Array Problem (LAP) in the sense that LAP can be reduced to TDCP. Since LAP has been shown to be NP-hard in the strong sense when the graph is of tree structure, TDCP is NP-hard for tree-structured temporal distance constraints of arbitrary values. Han, Lin and Liu have thus shown that TDCP is NP-hard in the strong sense even for a set of unit-execution-time jobs having chain-structured temporal distance constraints of arbitrary values and arbitrary number of jobs in each chain. For temporal distance constraints of the same constant value, Han, Lin and Liu introduced an O(n2) algorithm (where n is the total number of jobs), an O(m2c2) algorithm (where m is the number of chains and c is the constant temporal distance constraint) for solving a special case of TDCP where the graph is of directed chain tree and an O(n log n) algorithm for solving the special bi-level case. In this paper, another special case of the bi-level chain tree is studied. The graph of the temporal distance constraints is restricted to a bi-level directed tree with at most two sons, and present an O(n2) algorithm ND.